Multiple values for @calendar in the same <date> element

classic Classic list List threaded Threaded
24 messages Options
12
Reply | Threaded
Open this post in threaded view
|

Multiple values for @calendar in the same <date> element

Sascha Grabsch
Dear list members,

I have a question regarding the encoding of calendar systems using
@calendar with <date>.

Suppose we have a date given in a manuscript like this:

> Dienstag 7/19 May

This date is given in two calendar system, Julian and Gregorian. Would
it be possible to tag it like this:

<date calendar="#julian-gregorian" when="1829-05-19">Dienstag 7/19
May</date>.

My colleagues and I are unsure if this would be a valid usage of the
<date> element, but I think this would be a clean way to mark up the
given date (which in my reading is a _single_ date) and note the fact
that it is given in two different calendar systems. If at all possible I
would like to avoid having two <date> elements or *gasp* having to nest
the <date> elements.

We also are not sure what the URIs of @calendar should be pointing to in
general. Are there defined concepts/authority file references available
for calendar systems? I could only find examples of `#gregorian`,
`#julian` in the TEI Guidelines.

I would be very interested in your opinions and - if available -
examples!

All the best,

Sascha Grabsch
Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

Kurz, Stephan
Hi Sascha,

the Guidelines suggest that you can provide a second @when-custom date on which you provide information through @datingMethod.

<date calendar="#gregorian" when="1829-05-19" when-custom="1829-05-07" datingMethod="#julian">Dienstag 7/19 May</date>.

What bothers me is that the source already gives both calendars (as opposed to scenarios where you have some non-gregorian calendar used in the source and you normalize to Gregorian as a service for the reader and the occasional calendar events API)... One could read this as if two dates were recorded in the source, but seem to be part of the same <dateline>, even though I see why you would want to avoid that...

Greetings, Stephan

-----Original Message-----
From: TEI (Text Encoding Initiative) public discussion list <[hidden email]> On Behalf Of Sascha Grabsch
Sent: Wednesday, August 26, 2020 12:59 PM
To: [hidden email]
Subject: Multiple values for @calendar in the same <date> element

Dear list members,

I have a question regarding the encoding of calendar systems using
@calendar with <date>.

Suppose we have a date given in a manuscript like this:

> Dienstag 7/19 May

This date is given in two calendar system, Julian and Gregorian. Would
it be possible to tag it like this:

<date calendar="#julian-gregorian" when="1829-05-19">Dienstag 7/19
May</date>.

My colleagues and I are unsure if this would be a valid usage of the
<date> element, but I think this would be a clean way to mark up the
given date (which in my reading is a _single_ date) and note the fact
that it is given in two different calendar systems. If at all possible I
would like to avoid having two <date> elements or *gasp* having to nest
the <date> elements.

We also are not sure what the URIs of @calendar should be pointing to in
general. Are there defined concepts/authority file references available
for calendar systems? I could only find examples of `#gregorian`,
`#julian` in the TEI Guidelines.

I would be very interested in your opinions and - if available -
examples!

All the best,

Sascha Grabsch
Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

Martin Holmes
Hi there,

I would define a special calendar for these dual-system dates, so:

<date calendar="#julian-gregorian" when="1829-05-19"
when-custom="1829-05-07" datingMethod="#julian">Dienstag 7/19 May</date>

This says that the original source is expressed in two calendars (you'd
have to create a calendar with explanation to point at in the header).

If you want to get more gnarly, you could actually do this:

<date calendar="#julian-gregorian"><date calendar="#julian"
dating-method="#julian" when-custom="1829-05-07">7</date>/<date
when="1829-05-19">19 May</date></date>

so that the relevant components of each date are tagged. You could make
your life even more complicated if you liked, by using spans to point to
the specific bits which are relevant for each date, so that you could
include the " May" bit in both date element contents.

But you can spend your whole life getting deeper and deeper into dates
and calendars...

Cheers,
Martin


On 2020-08-26 5:49 a.m., Kurz, Stephan wrote:

> Hi Sascha,
>
> the Guidelines suggest that you can provide a second @when-custom date on which you provide information through @datingMethod.
>
> <date calendar="#gregorian" when="1829-05-19" when-custom="1829-05-07" datingMethod="#julian">Dienstag 7/19 May</date>.
>
> What bothers me is that the source already gives both calendars (as opposed to scenarios where you have some non-gregorian calendar used in the source and you normalize to Gregorian as a service for the reader and the occasional calendar events API)... One could read this as if two dates were recorded in the source, but seem to be part of the same <dateline>, even though I see why you would want to avoid that...
>
> Greetings, Stephan
>
> -----Original Message-----
> From: TEI (Text Encoding Initiative) public discussion list <[hidden email]> On Behalf Of Sascha Grabsch
> Sent: Wednesday, August 26, 2020 12:59 PM
> To: [hidden email]
> Subject: Multiple values for @calendar in the same <date> element
>
> Dear list members,
>
> I have a question regarding the encoding of calendar systems using
> @calendar with <date>.
>
> Suppose we have a date given in a manuscript like this:
>
>> Dienstag 7/19 May
>
> This date is given in two calendar system, Julian and Gregorian. Would
> it be possible to tag it like this:
>
> <date calendar="#julian-gregorian" when="1829-05-19">Dienstag 7/19
> May</date>.
>
> My colleagues and I are unsure if this would be a valid usage of the
> <date> element, but I think this would be a clean way to mark up the
> given date (which in my reading is a _single_ date) and note the fact
> that it is given in two different calendar systems. If at all possible I
> would like to avoid having two <date> elements or *gasp* having to nest
> the <date> elements.
>
> We also are not sure what the URIs of @calendar should be pointing to in
> general. Are there defined concepts/authority file references available
> for calendar systems? I could only find examples of `#gregorian`,
> `#julian` in the TEI Guidelines.
>
> I would be very interested in your opinions and - if available -
> examples!
>
> All the best,
>
> Sascha Grabsch
>

--
-------------------------------------
Humanities Computing and Media Centre
University of Victoria
[hidden email]
lou
Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

lou
In reply to this post by Sascha Grabsch
The second part of your question is easy: the value of the @calendar attribute is a pointer to a definition of a calendar system. The definition might be provided by a <calendar> element inside a <calendarDesc> in your TEI header, or by some other location  entirely, like a Wikipedia entry.

The first part is trickier. Does the @calendar reference apply to the value of @when, or to the content of the element? Hands up if your first thought (like mine) was the former! You're wrong. There's a warning note at https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html which says clearly 
"the calendar attribute (unlike @datingMethod defined in att.datable.custom) defines the calendar system of the date in the original material defined by the parent element, not the calendar to which the date is normalized."
So, if you want to indicate that the content of your date element is neither Julian nor Gregorian, but a combination of the two, you will have to define a new calendar (let's call it "julGreg") and reference that:
<date calendar="#julGreg" when="1829-05-19">Dienstag 7/19
May</date> , exactly as you propose. Note that @when is always understood to be normalised to the same (Gregorian) calendar. If you want to supply an additional normalisation to the Julian calendar, you'd have to add further attributes:
<date calendar="#julGreg" when="1829-05-19" when-custom="1829-05-07" dating-method="#julian">Dienstag 7/19 May</date>

You will of course also need to provide somewhere a calendar element like this:
<calendar xml:id="julGreg"><p>A composite calendar format in which both Julian and Gregorian forms are provided, separated by a solidus, with the Julian form first.</calendar>



On Wed, 26 Aug 2020 at 12:09, Sascha Grabsch <[hidden email]> wrote:
Dear list members,

I have a question regarding the encoding of calendar systems using
@calendar with <date>.

Suppose we have a date given in a manuscript like this:

> Dienstag 7/19 May

This date is given in two calendar system, Julian and Gregorian. Would
it be possible to tag it like this:

<date calendar="#julian-gregorian" when="1829-05-19">Dienstag 7/19
May</date>.

My colleagues and I are unsure if this would be a valid usage of the
<date> element, but I think this would be a clean way to mark up the
given date (which in my reading is a _single_ date) and note the fact
that it is given in two different calendar systems. If at all possible I
would like to avoid having two <date> elements or *gasp* having to nest
the <date> elements.

We also are not sure what the URIs of @calendar should be pointing to in
general. Are there defined concepts/authority file references available
for calendar systems? I could only find examples of `#gregorian`,
`#julian` in the TEI Guidelines.

I would be very interested in your opinions and - if available -
examples!

All the best,

Sascha Grabsch
Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

Christian Thomas (HU Berlin)

Dear TEI list, please excuse my maybe naive question, but, taking up the example from https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html which says:

He was born on <date calendar="#gregorian">Feb. 22, 1732</date> (<date calendar="#julianwhen="1732-02-22"> Feb. 11, 1731/32, O.S.</date>).

Why can't I join the two (or more) attribute values in @calendar like this:

He was born on <date calendar="#gregorian #julianwhen="1732-02-22">Feb. 22, 1732 (Feb. 11, 1731/32, O.S.</date>).

This would (in my reading) say that in the source both calendars are being used, @when is the standardised Gregorian date as specified in the TEI Guidelines, and would save us the trouble of having to invent multiple new values for 'combined' calendars.

Please excuse me if this is layed out as impossible resp. forbidden somewhere else (and, if so, thanks for pointing me to that place).

Thanks, best wishes
Christian




Am 26.08.2020 um 17:19 schrieb Lou Burnard:
The second part of your question is easy: the value of the @calendar attribute is a pointer to a definition of a calendar system. The definition might be provided by a <calendar> element inside a <calendarDesc> in your TEI header, or by some other location  entirely, like a Wikipedia entry.

The first part is trickier. Does the @calendar reference apply to the value of @when, or to the content of the element? Hands up if your first thought (like mine) was the former! You're wrong. There's a warning note at https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html which says clearly 
"the calendar attribute (unlike @datingMethod defined in att.datable.custom) defines the calendar system of the date in the original material defined by the parent element, not the calendar to which the date is normalized."
So, if you want to indicate that the content of your date element is neither Julian nor Gregorian, but a combination of the two, you will have to define a new calendar (let's call it "julGreg") and reference that:
<date calendar="#julGreg" when="1829-05-19">Dienstag 7/19
May</date> , exactly as you propose. Note that @when is always understood to be normalised to the same (Gregorian) calendar. If you want to supply an additional normalisation to the Julian calendar, you'd have to add further attributes:
<date calendar="#julGreg" when="1829-05-19" when-custom="1829-05-07" dating-method="#julian">Dienstag 7/19 May</date>

You will of course also need to provide somewhere a calendar element like this:
<calendar xml:id="julGreg"><p>A composite calendar format in which both Julian and Gregorian forms are provided, separated by a solidus, with the Julian form first.</calendar>



On Wed, 26 Aug 2020 at 12:09, Sascha Grabsch <[hidden email]> wrote:
Dear list members,

I have a question regarding the encoding of calendar systems using
@calendar with <date>.

Suppose we have a date given in a manuscript like this:

> Dienstag 7/19 May

This date is given in two calendar system, Julian and Gregorian. Would
it be possible to tag it like this:

<date calendar="#julian-gregorian" when="1829-05-19">Dienstag 7/19
May</date>.

My colleagues and I are unsure if this would be a valid usage of the
<date> element, but I think this would be a clean way to mark up the
given date (which in my reading is a _single_ date) and note the fact
that it is given in two different calendar systems. If at all possible I
would like to avoid having two <date> elements or *gasp* having to nest
the <date> elements.

We also are not sure what the URIs of @calendar should be pointing to in
general. Are there defined concepts/authority file references available
for calendar systems? I could only find examples of `#gregorian`,
`#julian` in the TEI Guidelines.

I would be very interested in your opinions and - if available -
examples!

All the best,

Sascha Grabsch

-- 
Christian Thomas
[hidden email]
@dta_cthomas
--
Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

Martin Holmes
Unfortunately, @calendar currently allows only one pointer:

<https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>

but I think you should be able to use multiple pointers in this case;
the use-case is very straightforward. Can anyone think of a reason why not?

The same is not true of @datingMethod, though; because that specifies
what system is being used for other attributes, and it's not possible to
say which of the array of @*-custom attributes would use one dating
method and which another, @datingMethod should stay as a single pointer.

Cheers,
Martin

On 2020-08-26 9:20 a.m., Christian Thomas wrote:

>
> Dear TEI list, please excuse my maybe naive question, but, taking up the
> example from
> https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html 
> which says:
>
>> He was born on <date calendar="#gregorian">Feb. 22, 1732</date>(<date
>> calendar="#julian" when="1732-02-22">Feb. 11, 1731/32, O.S.</date>).
>
> Why can't I join the two (or more) attribute values in @calendar like this:
>
>> He was born on <date calendar="#gregorian#julian"
>> when="1732-02-22">Feb. 22, 1732(Feb. 11, 1731/32, O.S.</date>).
>
> This would (in my reading) say that in the source both calendars are
> being used, @when is the standardised Gregorian date as specified in the
> TEI Guidelines, and would save us the trouble of having to invent
> multiple new values for 'combined' calendars.
>
> Please excuse me if this is layed out as impossible resp. forbidden
> somewhere else (and, if so, thanks for pointing me to that place).
>
> Thanks, best wishes
> Christian
>
>
>
>
> Am 26.08.2020 um 17:19 schrieb Lou Burnard:
>> The second part of your question is easy: the value of the @calendar
>> attribute is a pointer to a definition of a calendar system. The
>> definition might be provided by a <calendar> element inside a
>> <calendarDesc> in your TEI header, or by some other location  
>> entirely, like a Wikipedia entry.
>>
>> The first part is trickier. Does the @calendar reference apply to the
>> value of @when, or to the content of the element? Hands up if your
>> first thought (like mine) was the former! You're wrong. There's a
>> warning note at
>> https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html 
>> which says clearly
>> "the calendar attribute (unlike @datingMethod defined in
>> att.datable.custom
>> <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>)
>> defines the calendar system of the date in the original material
>> defined by the parent element, /not/ the calendar to which the date is
>> normalized."
>> So, if you want to indicate that the content of your date element is
>> neither Julian nor Gregorian, but a combination of the two, you will
>> have to define a new calendar (let's call it "julGreg") and reference
>> that:
>> <date calendar="#julGreg" when="1829-05-19">Dienstag 7/19
>> May</date> , exactly as you propose. Note that @when is always
>> understood to be normalised to the same (Gregorian) calendar. If you
>> want to supply an additional normalisation to the Julian calendar,
>> you'd have to add further attributes:
>> <date calendar="#julGreg" when="1829-05-19" when-custom="1829-05-07"
>> dating-method="#julian">Dienstag 7/19 May</date>
>>
>> You will of course also need to provide somewhere a calendar element
>> like this:
>> <calendar xml:id="julGreg"><p>A composite calendar format in which
>> both Julian and Gregorian forms are provided, separated by a solidus,
>> with the Julian form first.</calendar>
>>
>>
>>
>> On Wed, 26 Aug 2020 at 12:09, Sascha Grabsch <[hidden email]
>> <mailto:[hidden email]>> wrote:
>>
>>     Dear list members,
>>
>>     I have a question regarding the encoding of calendar systems using
>>     @calendar with <date>.
>>
>>     Suppose we have a date given in a manuscript like this:
>>
>>     > Dienstag 7/19 May
>>
>>     This date is given in two calendar system, Julian and Gregorian.
>>     Would
>>     it be possible to tag it like this:
>>
>>     <date calendar="#julian-gregorian" when="1829-05-19">Dienstag 7/19
>>     May</date>.
>>
>>     My colleagues and I are unsure if this would be a valid usage of the
>>     <date> element, but I think this would be a clean way to mark up the
>>     given date (which in my reading is a _single_ date) and note the fact
>>     that it is given in two different calendar systems. If at all
>>     possible I
>>     would like to avoid having two <date> elements or *gasp* having to
>>     nest
>>     the <date> elements.
>>
>>     We also are not sure what the URIs of @calendar should be pointing
>>     to in
>>     general. Are there defined concepts/authority file references
>>     available
>>     for calendar systems? I could only find examples of `#gregorian`,
>>     `#julian` in the TEI Guidelines.
>>
>>     I would be very interested in your opinions and - if available -
>>     examples!
>>
>>     All the best,
>>
>>     Sascha Grabsch
>>
>
> --
> Christian Thomas
> [hidden email]
> @dta_cthomas
> --
>

--
-------------------------------------
Humanities Computing and Media Centre
University of Victoria
[hidden email]
Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

Christian Thomas (HU Berlin)
Dear Martin, dear all, as you can guess:

Am 26.08.2020 um 18:28 schrieb Martin Holmes:
> Unfortunately, @calendar currently allows only one pointer:
>
> <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
>
> but I think you should be able to use multiple pointers in this case;
> the use-case is very straightforward. Can anyone think of a reason why
> not?

I would be very much in favour of this! This, on the other hand:

> The same is not true of @datingMethod, though;

is a different matter and absolutely reasonable. There can be only one
@datingMethod, but (obviously) more than one @calendar used in the
source text.

Best wishes
Christian


> because that specifies what system is being used for other attributes,
> and it's not possible to say which of the array of @*-custom
> attributes would use one dating method and which another,
> @datingMethod should stay as a single pointer.
>
> Cheers,
> Martin
>
> On 2020-08-26 9:20 a.m., Christian Thomas wrote:
>>
>> Dear TEI list, please excuse my maybe naive question, but, taking up
>> the example from
>> https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html 
>> which says:
>>
>>> He was born on <date calendar="#gregorian">Feb. 22,
>>> 1732</date>(<date calendar="#julian" when="1732-02-22">Feb. 11,
>>> 1731/32, O.S.</date>).
>>
>> Why can't I join the two (or more) attribute values in @calendar like
>> this:
>>
>>> He was born on <date calendar="#gregorian#julian"
>>> when="1732-02-22">Feb. 22, 1732(Feb. 11, 1731/32, O.S.</date>).
>>
>> This would (in my reading) say that in the source both calendars are
>> being used, @when is the standardised Gregorian date as specified in
>> the TEI Guidelines, and would save us the trouble of having to invent
>> multiple new values for 'combined' calendars.
>>
>> Please excuse me if this is layed out as impossible resp. forbidden
>> somewhere else (and, if so, thanks for pointing me to that place).
>>
>> Thanks, best wishes
>> Christian
>>
>>
>>
>>
>> Am 26.08.2020 um 17:19 schrieb Lou Burnard:
>>> The second part of your question is easy: the value of the @calendar
>>> attribute is a pointer to a definition of a calendar system. The
>>> definition might be provided by a <calendar> element inside a
>>> <calendarDesc> in your TEI header, or by some other location 
>>> entirely, like a Wikipedia entry.
>>>
>>> The first part is trickier. Does the @calendar reference apply to
>>> the value of @when, or to the content of the element? Hands up if
>>> your first thought (like mine) was the former! You're wrong. There's
>>> a warning note at
>>> https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html 
>>> which says clearly
>>> "the calendar attribute (unlike @datingMethod defined in
>>> att.datable.custom
>>> <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>)
>>> defines the calendar system of the date in the original material
>>> defined by the parent element, /not/ the calendar to which the date
>>> is normalized."
>>> So, if you want to indicate that the content of your date element is
>>> neither Julian nor Gregorian, but a combination of the two, you will
>>> have to define a new calendar (let's call it "julGreg") and
>>> reference that:
>>> <date calendar="#julGreg" when="1829-05-19">Dienstag 7/19
>>> May</date> , exactly as you propose. Note that @when is always
>>> understood to be normalised to the same (Gregorian) calendar. If you
>>> want to supply an additional normalisation to the Julian calendar,
>>> you'd have to add further attributes:
>>> <date calendar="#julGreg" when="1829-05-19" when-custom="1829-05-07"
>>> dating-method="#julian">Dienstag 7/19 May</date>
>>>
>>> You will of course also need to provide somewhere a calendar element
>>> like this:
>>> <calendar xml:id="julGreg"><p>A composite calendar format in which
>>> both Julian and Gregorian forms are provided, separated by a
>>> solidus, with the Julian form first.</calendar>
>>>
>>>
>>>
>>> On Wed, 26 Aug 2020 at 12:09, Sascha Grabsch <[hidden email]
>>> <mailto:[hidden email]>> wrote:
>>>
>>>     Dear list members,
>>>
>>>     I have a question regarding the encoding of calendar systems using
>>>     @calendar with <date>.
>>>
>>>     Suppose we have a date given in a manuscript like this:
>>>
>>>     > Dienstag 7/19 May
>>>
>>>     This date is given in two calendar system, Julian and Gregorian.
>>>     Would
>>>     it be possible to tag it like this:
>>>
>>>     <date calendar="#julian-gregorian" when="1829-05-19">Dienstag 7/19
>>>     May</date>.
>>>
>>>     My colleagues and I are unsure if this would be a valid usage of
>>> the
>>>     <date> element, but I think this would be a clean way to mark up
>>> the
>>>     given date (which in my reading is a _single_ date) and note the
>>> fact
>>>     that it is given in two different calendar systems. If at all
>>>     possible I
>>>     would like to avoid having two <date> elements or *gasp* having to
>>>     nest
>>>     the <date> elements.
>>>
>>>     We also are not sure what the URIs of @calendar should be pointing
>>>     to in
>>>     general. Are there defined concepts/authority file references
>>>     available
>>>     for calendar systems? I could only find examples of `#gregorian`,
>>>     `#julian` in the TEI Guidelines.
>>>
>>>     I would be very interested in your opinions and - if available -
>>>     examples!
>>>
>>>     All the best,
>>>
>>>     Sascha Grabsch
>>>
>>
>> --
>> Christian Thomas
>> [hidden email]
>> @dta_cthomas
>> --
>>
>

--
Christian Thomas
[hidden email]
@dta_cthomas
--
Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

Christian Thomas (HU Berlin)
In reply to this post by Martin Holmes

Hi again, just to correct my own mistakes, I don't know where all the whitespaces went:
> He was born on <date calendar="#gregorian #julian" when="1732-02-22">Feb. 22, 1732 (Feb. 11, 1731/32, O.S.</date>).
Or, to take up Sascha Grabschs initial expample from the edition humboldt digital (cf. https://edition-humboldt.de/H0016785/12r, currently tagged wrong, alas the question), "Dienstag 7/19 May" should be
> <date when="1829-05-19" calendar="#gregorian #julian">Dienstag <hi rendition="#sup">7</hi>/<hi rendition="#sub">19</hi> May</date>

All in favour?

Best wishes
Christian


Am 26.08.2020 um 18:28 schrieb Martin Holmes:
Unfortunately, @calendar currently allows only one pointer:

<https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>

but I think you should be able to use multiple pointers in this case; the use-case is very straightforward. Can anyone think of a reason why not?

The same is not true of @datingMethod, though; because that specifies what system is being used for other attributes, and it's not possible to say which of the array of @*-custom attributes would use one dating method and which another, @datingMethod should stay as a single pointer.

Cheers,
Martin

On 2020-08-26 9:20 a.m., Christian Thomas wrote:

Dear TEI list, please excuse my maybe naive question, but, taking up the example from https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html which says:

He was born on <date calendar="#gregorian">Feb. 22, 1732</date>(<date calendar="#julian" when="1732-02-22">Feb. 11, 1731/32, O.S.</date>).

Why can't I join the two (or more) attribute values in @calendar like this:

He was born on <date calendar="#gregorian#julian" when="1732-02-22">Feb. 22, 1732(Feb. 11, 1731/32, O.S.</date>).

This would (in my reading) say that in the source both calendars are being used, @when is the standardised Gregorian date as specified in the TEI Guidelines, and would save us the trouble of having to invent multiple new values for 'combined' calendars.

Please excuse me if this is layed out as impossible resp. forbidden somewhere else (and, if so, thanks for pointing me to that place).

Thanks, best wishes
Christian




Am 26.08.2020 um 17:19 schrieb Lou Burnard:
The second part of your question is easy: the value of the @calendar attribute is a pointer to a definition of a calendar system. The definition might be provided by a <calendar> element inside a <calendarDesc> in your TEI header, or by some other location  entirely, like a Wikipedia entry.

The first part is trickier. Does the @calendar reference apply to the value of @when, or to the content of the element? Hands up if your first thought (like mine) was the former! You're wrong. There's a warning note at https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html which says clearly
"the calendar attribute (unlike @datingMethod defined in att.datable.custom <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>) defines the calendar system of the date in the original material defined by the parent element, /not/ the calendar to which the date is normalized."
So, if you want to indicate that the content of your date element is neither Julian nor Gregorian, but a combination of the two, you will have to define a new calendar (let's call it "julGreg") and reference that:
<date calendar="#julGreg" when="1829-05-19">Dienstag 7/19
May</date> , exactly as you propose. Note that @when is always understood to be normalised to the same (Gregorian) calendar. If you want to supply an additional normalisation to the Julian calendar, you'd have to add further attributes:
<date calendar="#julGreg" when="1829-05-19" when-custom="1829-05-07" dating-method="#julian">Dienstag 7/19 May</date>

You will of course also need to provide somewhere a calendar element like this:
<calendar xml:id="julGreg"><p>A composite calendar format in which both Julian and Gregorian forms are provided, separated by a solidus, with the Julian form first.</calendar>



On Wed, 26 Aug 2020 at 12:09, Sascha Grabsch <[hidden email] [hidden email]> wrote:

    Dear list members,

    I have a question regarding the encoding of calendar systems using
    @calendar with <date>.

    Suppose we have a date given in a manuscript like this:

    > Dienstag 7/19 May

    This date is given in two calendar system, Julian and Gregorian.
    Would
    it be possible to tag it like this:

    <date calendar="#julian-gregorian" when="1829-05-19">Dienstag 7/19
    May</date>.

    My colleagues and I are unsure if this would be a valid usage of the
    <date> element, but I think this would be a clean way to mark up the
    given date (which in my reading is a _single_ date) and note the fact
    that it is given in two different calendar systems. If at all
    possible I
    would like to avoid having two <date> elements or *gasp* having to
    nest
    the <date> elements.

    We also are not sure what the URIs of @calendar should be pointing
    to in
    general. Are there defined concepts/authority file references
    available
    for calendar systems? I could only find examples of `#gregorian`,
    `#julian` in the TEI Guidelines.

    I would be very interested in your opinions and - if available -
    examples!

    All the best,

    Sascha Grabsch


-- 
Christian Thomas
[hidden email]
@dta_cthomas
--



-- 
Christian Thomas
[hidden email]
@dta_cthomas
--
lou
Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

lou
I am not sure what you are trying to achieve here. The calendar attribute permits only a single Uri. A Uri may begin with but not contain the # character. So whether or not you put spaces into it your suggested values are syntactically invalid.

On Wed, 26 Aug 2020, 19:13 Christian Thomas, <[hidden email]> wrote:

Hi again, just to correct my own mistakes, I don't know where all the whitespaces went:
> He was born on <date calendar="#gregorian #julian" when="1732-02-22">Feb. 22, 1732 (Feb. 11, 1731/32, O.S.</date>).
Or, to take up Sascha Grabschs initial expample from the edition humboldt digital (cf. https://edition-humboldt.de/H0016785/12r, currently tagged wrong, alas the question), "Dienstag 7/19 May" should be
> <date when="1829-05-19" calendar="#gregorian #julian">Dienstag <hi rendition="#sup">7</hi>/<hi rendition="#sub">19</hi> May</date>

All in favour?

Best wishes
Christian


Am 26.08.2020 um 18:28 schrieb Martin Holmes:
Unfortunately, @calendar currently allows only one pointer:

<https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>

but I think you should be able to use multiple pointers in this case; the use-case is very straightforward. Can anyone think of a reason why not?

The same is not true of @datingMethod, though; because that specifies what system is being used for other attributes, and it's not possible to say which of the array of @*-custom attributes would use one dating method and which another, @datingMethod should stay as a single pointer.

Cheers,
Martin

On 2020-08-26 9:20 a.m., Christian Thomas wrote:

Dear TEI list, please excuse my maybe naive question, but, taking up the example from https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html which says:

He was born on <date calendar="#gregorian">Feb. 22, 1732</date>(<date calendar="#julian" when="1732-02-22">Feb. 11, 1731/32, O.S.</date>).

Why can't I join the two (or more) attribute values in @calendar like this:

He was born on <date calendar="#gregorian#julian" when="1732-02-22">Feb. 22, 1732(Feb. 11, 1731/32, O.S.</date>).

This would (in my reading) say that in the source both calendars are being used, @when is the standardised Gregorian date as specified in the TEI Guidelines, and would save us the trouble of having to invent multiple new values for 'combined' calendars.

Please excuse me if this is layed out as impossible resp. forbidden somewhere else (and, if so, thanks for pointing me to that place).

Thanks, best wishes
Christian




Am 26.08.2020 um 17:19 schrieb Lou Burnard:
The second part of your question is easy: the value of the @calendar attribute is a pointer to a definition of a calendar system. The definition might be provided by a <calendar> element inside a <calendarDesc> in your TEI header, or by some other location  entirely, like a Wikipedia entry.

The first part is trickier. Does the @calendar reference apply to the value of @when, or to the content of the element? Hands up if your first thought (like mine) was the former! You're wrong. There's a warning note at https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html which says clearly
"the calendar attribute (unlike @datingMethod defined in att.datable.custom <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>) defines the calendar system of the date in the original material defined by the parent element, /not/ the calendar to which the date is normalized."
So, if you want to indicate that the content of your date element is neither Julian nor Gregorian, but a combination of the two, you will have to define a new calendar (let's call it "julGreg") and reference that:
<date calendar="#julGreg" when="1829-05-19">Dienstag 7/19
May</date> , exactly as you propose. Note that @when is always understood to be normalised to the same (Gregorian) calendar. If you want to supply an additional normalisation to the Julian calendar, you'd have to add further attributes:
<date calendar="#julGreg" when="1829-05-19" when-custom="1829-05-07" dating-method="#julian">Dienstag 7/19 May</date>

You will of course also need to provide somewhere a calendar element like this:
<calendar xml:id="julGreg"><p>A composite calendar format in which both Julian and Gregorian forms are provided, separated by a solidus, with the Julian form first.</calendar>



On Wed, 26 Aug 2020 at 12:09, Sascha Grabsch <[hidden email] [hidden email]> wrote:

    Dear list members,

    I have a question regarding the encoding of calendar systems using
    @calendar with <date>.

    Suppose we have a date given in a manuscript like this:

    > Dienstag 7/19 May

    This date is given in two calendar system, Julian and Gregorian.
    Would
    it be possible to tag it like this:

    <date calendar="#julian-gregorian" when="1829-05-19">Dienstag 7/19
    May</date>.

    My colleagues and I are unsure if this would be a valid usage of the
    <date> element, but I think this would be a clean way to mark up the
    given date (which in my reading is a _single_ date) and note the fact
    that it is given in two different calendar systems. If at all
    possible I
    would like to avoid having two <date> elements or *gasp* having to
    nest
    the <date> elements.

    We also are not sure what the URIs of @calendar should be pointing
    to in
    general. Are there defined concepts/authority file references
    available
    for calendar systems? I could only find examples of `#gregorian`,
    `#julian` in the TEI Guidelines.

    I would be very interested in your opinions and - if available -
    examples!

    All the best,

    Sascha Grabsch


-- 
Christian Thomas
[hidden email]
@dta_cthomas
--



-- 
Christian Thomas
[hidden email]
@dta_cthomas
--
Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

Martin Holmes
Hi Lou,

I think Christian was suggesting that @calendar permit multiple
pointers, which I think makes perfect sense for scenarios like this. If
you know of any good reasons not to make such a feature request, let us
know; I think it would be a good idea myself.

Cheers,
Martin

On 2020-08-26 12:20 p.m., Lou Burnard wrote:

> I am not sure what you are trying to achieve here. The calendar
> attribute permits only a single Uri. A Uri may begin with but not
> contain the # character. So whether or not you put spaces into it your
> suggested values are syntactically invalid.
>
> On Wed, 26 Aug 2020, 19:13 Christian Thomas,
> <[hidden email]
> <mailto:[hidden email]>> wrote:
>
>
>     Hi again, just to correct my own mistakes, I don't know where all
>     the whitespaces went:
>
>     > He was born on <date calendar="#gregorian #julian" when="1732-02-22">Feb. 22, 1732 (Feb. 11, 1731/32, O.S.</date>).
>
>     Or, to take up Sascha Grabschs initial expample from the /edition
>     humboldt digital/ (cf. https://edition-humboldt.de/H0016785/12r,
>     currently tagged wrong, alas the question), "Dienstag 7/19 May"
>     should be
>
>     > <date when="1829-05-19" calendar="#gregorian #julian">Dienstag <hi rendition="#sup">7</hi>/<hi rendition="#sub">19</hi> May</date>
>
>
>     All in favour?
>
>     Best wishes
>     Christian
>
>
>     Am 26.08.2020 um 18:28 schrieb Martin Holmes:
>>     Unfortunately, @calendar currently allows only one pointer:
>>
>>     <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
>>     <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
>>
>>
>>     but I think you should be able to use multiple pointers in this
>>     case; the use-case is very straightforward. Can anyone think of a
>>     reason why not?
>>
>>     The same is not true of @datingMethod, though; because that
>>     specifies what system is being used for other attributes, and it's
>>     not possible to say which of the array of @*-custom attributes
>>     would use one dating method and which another, @datingMethod
>>     should stay as a single pointer.
>>
>>     Cheers,
>>     Martin
>>
>>     On 2020-08-26 9:20 a.m., Christian Thomas wrote:
>>>
>>>     Dear TEI list, please excuse my maybe naive question, but, taking
>>>     up the example from
>>>     https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
>>>     which says:
>>>
>>>>     He was born on <date calendar="#gregorian">Feb. 22,
>>>>     1732</date>(<date calendar="#julian" when="1732-02-22">Feb. 11,
>>>>     1731/32, O.S.</date>).
>>>
>>>     Why can't I join the two (or more) attribute values in @calendar
>>>     like this:
>>>
>>>>     He was born on <date calendar="#gregorian#julian"
>>>>     when="1732-02-22">Feb. 22, 1732(Feb. 11, 1731/32, O.S.</date>).
>>>
>>>     This would (in my reading) say that in the source both calendars
>>>     are being used, @when is the standardised Gregorian date as
>>>     specified in the TEI Guidelines, and would save us the trouble of
>>>     having to invent multiple new values for 'combined' calendars.
>>>
>>>     Please excuse me if this is layed out as impossible resp.
>>>     forbidden somewhere else (and, if so, thanks for pointing me to
>>>     that place).
>>>
>>>     Thanks, best wishes
>>>     Christian
>>>
>>>
>>>
>>>
>>>     Am 26.08.2020 um 17:19 schrieb Lou Burnard:
>>>>     The second part of your question is easy: the value of the
>>>>     @calendar attribute is a pointer to a definition of a calendar
>>>>     system. The definition might be provided by a <calendar> element
>>>>     inside a <calendarDesc> in your TEI header, or by some other
>>>>     location  entirely, like a Wikipedia entry.
>>>>
>>>>     The first part is trickier. Does the @calendar reference apply
>>>>     to the value of @when, or to the content of the element? Hands
>>>>     up if your first thought (like mine) was the former! You're
>>>>     wrong. There's a warning note at
>>>>     https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
>>>>     which says clearly
>>>>     "the calendar attribute (unlike @datingMethod defined in
>>>>     att.datable.custom
>>>>     <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>
>>>>     <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>)
>>>>     defines the calendar system of the date in the original material
>>>>     defined by the parent element, /not/ the calendar to which the
>>>>     date is normalized."
>>>>     So, if you want to indicate that the content of your date
>>>>     element is neither Julian nor Gregorian, but a combination of
>>>>     the two, you will have to define a new calendar (let's call it
>>>>     "julGreg") and reference that:
>>>>     <date calendar="#julGreg" when="1829-05-19">Dienstag 7/19
>>>>     May</date> , exactly as you propose. Note that @when is always
>>>>     understood to be normalised to the same (Gregorian) calendar. If
>>>>     you want to supply an additional normalisation to the Julian
>>>>     calendar, you'd have to add further attributes:
>>>>     <date calendar="#julGreg" when="1829-05-19"
>>>>     when-custom="1829-05-07" dating-method="#julian">Dienstag 7/19
>>>>     May</date>
>>>>
>>>>     You will of course also need to provide somewhere a calendar
>>>>     element like this:
>>>>     <calendar xml:id="julGreg"><p>A composite calendar format in
>>>>     which both Julian and Gregorian forms are provided, separated by
>>>>     a solidus, with the Julian form first.</calendar>
>>>>
>>>>
>>>>
>>>>     On Wed, 26 Aug 2020 at 12:09, Sascha Grabsch <[hidden email]
>>>>     <mailto:[hidden email]> <mailto:[hidden email]>
>>>>     <mailto:[hidden email]>> wrote:
>>>>
>>>>         Dear list members,
>>>>
>>>>         I have a question regarding the encoding of calendar systems
>>>>     using
>>>>         @calendar with <date>.
>>>>
>>>>         Suppose we have a date given in a manuscript like this:
>>>>
>>>>         > Dienstag 7/19 May
>>>>
>>>>         This date is given in two calendar system, Julian and
>>>>     Gregorian.
>>>>         Would
>>>>         it be possible to tag it like this:
>>>>
>>>>         <date calendar="#julian-gregorian"
>>>>     when="1829-05-19">Dienstag 7/19
>>>>         May</date>.
>>>>
>>>>         My colleagues and I are unsure if this would be a valid
>>>>     usage of the
>>>>         <date> element, but I think this would be a clean way to
>>>>     mark up the
>>>>         given date (which in my reading is a _single_ date) and note
>>>>     the fact
>>>>         that it is given in two different calendar systems. If at all
>>>>         possible I
>>>>         would like to avoid having two <date> elements or *gasp*
>>>>     having to
>>>>         nest
>>>>         the <date> elements.
>>>>
>>>>         We also are not sure what the URIs of @calendar should be
>>>>     pointing
>>>>         to in
>>>>         general. Are there defined concepts/authority file references
>>>>         available
>>>>         for calendar systems? I could only find examples of
>>>>     `#gregorian`,
>>>>         `#julian` in the TEI Guidelines.
>>>>
>>>>         I would be very interested in your opinions and - if
>>>>     available -
>>>>         examples!
>>>>
>>>>         All the best,
>>>>
>>>>         Sascha Grabsch
>>>>
>>>
>>>     --
>>>     Christian Thomas
>>>     [hidden email]
>>>     <mailto:[hidden email]>
>>>     @dta_cthomas
>>>     --
>>>
>>
>
>     --
>     Christian Thomas
>     [hidden email]  <mailto:[hidden email]>
>     @dta_cthomas
>     --
>

--
-------------------------------------
Humanities Computing and Media Centre
University of Victoria
[hidden email]
Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

Paul Schaffner
Am I right in understanding this request to be aimed specifically
and exclusively at representations of date that more-or-less
inextricably combine references to more than one calendar? Hence

 <date when="1829-05-19" calendar="#gregorian #julian">Thursday, 7/19 May</date>
 
... whereas where the expression is separable, the need goes away?
 
 <date when="1829-05-19" calendar="#julian">Thursday, 7 May, O.S.</date> or
 <date when="1829-05-19" calendar="#gregorian">Thursday, 19 May, N.S.</date>

... and that the only alternative to this suggestion available within
the schema currently is to create a 'calendar' that is in effect a hybrid
between those referenced? like this?

 <date when="1829-05-19" calendar="#julian_gregorian">Thursday, 7/19 May</date>

pfs
 
 

On Wed, Aug 26, 2020, at 16:23, Martin Holmes wrote:

> Hi Lou,
>
> I think Christian was suggesting that @calendar permit multiple
> pointers, which I think makes perfect sense for scenarios like this. If
> you know of any good reasons not to make such a feature request, let us
> know; I think it would be a good idea myself.
>
> Cheers,
> Martin
>
> On 2020-08-26 12:20 p.m., Lou Burnard wrote:
> > I am not sure what you are trying to achieve here. The calendar
> > attribute permits only a single Uri. A Uri may begin with but not
> > contain the # character. So whether or not you put spaces into it your
> > suggested values are syntactically invalid.
> >
> > On Wed, 26 Aug 2020, 19:13 Christian Thomas,
> > <[hidden email]
> > <mailto:[hidden email]>> wrote:
> >
> >
> >     Hi again, just to correct my own mistakes, I don't know where all
> >     the whitespaces went:
> >
> >     > He was born on <date calendar="#gregorian #julian" when="1732-02-22">Feb. 22, 1732 (Feb. 11, 1731/32, O.S.</date>).
> >
> >     Or, to take up Sascha Grabschs initial expample from the /edition
> >     humboldt digital/ (cf. https://edition-humboldt.de/H0016785/12r,
> >     currently tagged wrong, alas the question), "Dienstag 7/19 May"
> >     should be
> >
> >     > <date when="1829-05-19" calendar="#gregorian #julian">Dienstag <hi rendition="#sup">7</hi>/<hi rendition="#sub">19</hi> May</date>
> >
> >
> >     All in favour?
> >
> >     Best wishes
> >     Christian
> >
> >
> >     Am 26.08.2020 um 18:28 schrieb Martin Holmes:
> >>     Unfortunately, @calendar currently allows only one pointer:
> >>
> >>     <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
> >>     <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
> >>
> >>
> >>     but I think you should be able to use multiple pointers in this
> >>     case; the use-case is very straightforward. Can anyone think of a
> >>     reason why not?
> >>
> >>     The same is not true of @datingMethod, though; because that
> >>     specifies what system is being used for other attributes, and it's
> >>     not possible to say which of the array of @*-custom attributes
> >>     would use one dating method and which another, @datingMethod
> >>     should stay as a single pointer.
> >>
> >>     Cheers,
> >>     Martin
> >>
> >>     On 2020-08-26 9:20 a.m., Christian Thomas wrote:
> >>>
> >>>     Dear TEI list, please excuse my maybe naive question, but, taking
> >>>     up the example from
> >>>     https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
> >>>     which says:
> >>>
> >>>>     He was born on <date calendar="#gregorian">Feb. 22,
> >>>>     1732</date>(<date calendar="#julian" when="1732-02-22">Feb. 11,
> >>>>     1731/32, O.S.</date>).
> >>>
> >>>     Why can't I join the two (or more) attribute values in @calendar
> >>>     like this:
> >>>
> >>>>     He was born on <date calendar="#gregorian#julian"
> >>>>     when="1732-02-22">Feb. 22, 1732(Feb. 11, 1731/32, O.S.</date>).
> >>>
> >>>     This would (in my reading) say that in the source both calendars
> >>>     are being used, @when is the standardised Gregorian date as
> >>>     specified in the TEI Guidelines, and would save us the trouble of
> >>>     having to invent multiple new values for 'combined' calendars.
> >>>
> >>>     Please excuse me if this is layed out as impossible resp.
> >>>     forbidden somewhere else (and, if so, thanks for pointing me to
> >>>     that place).
> >>>
> >>>     Thanks, best wishes
> >>>     Christian
> >>>
> >>>
> >>>
> >>>
> >>>     Am 26.08.2020 um 17:19 schrieb Lou Burnard:
> >>>>     The second part of your question is easy: the value of the
> >>>>     @calendar attribute is a pointer to a definition of a calendar
> >>>>     system. The definition might be provided by a <calendar> element
> >>>>     inside a <calendarDesc> in your TEI header, or by some other
> >>>>     location  entirely, like a Wikipedia entry.
> >>>>
> >>>>     The first part is trickier. Does the @calendar reference apply
> >>>>     to the value of @when, or to the content of the element? Hands
> >>>>     up if your first thought (like mine) was the former! You're
> >>>>     wrong. There's a warning note at
> >>>>     https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
> >>>>     which says clearly
> >>>>     "the calendar attribute (unlike @datingMethod defined in
> >>>>     att.datable.custom
> >>>>     <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>
> >>>>     <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>)
> >>>>     defines the calendar system of the date in the original material
> >>>>     defined by the parent element, /not/ the calendar to which the
> >>>>     date is normalized."
> >>>>     So, if you want to indicate that the content of your date
> >>>>     element is neither Julian nor Gregorian, but a combination of
> >>>>     the two, you will have to define a new calendar (let's call it
> >>>>     "julGreg") and reference that:
> >>>>     <date calendar="#julGreg" when="1829-05-19">Dienstag 7/19
> >>>>     May</date> , exactly as you propose. Note that @when is always
> >>>>     understood to be normalised to the same (Gregorian) calendar. If
> >>>>     you want to supply an additional normalisation to the Julian
> >>>>     calendar, you'd have to add further attributes:
> >>>>     <date calendar="#julGreg" when="1829-05-19"
> >>>>     when-custom="1829-05-07" dating-method="#julian">Dienstag 7/19
> >>>>     May</date>
> >>>>
> >>>>     You will of course also need to provide somewhere a calendar
> >>>>     element like this:
> >>>>     <calendar xml:id="julGreg"><p>A composite calendar format in
> >>>>     which both Julian and Gregorian forms are provided, separated by
> >>>>     a solidus, with the Julian form first.</calendar>
> >>>>
> >>>>
> >>>>
> >>>>     On Wed, 26 Aug 2020 at 12:09, Sascha Grabsch <[hidden email]
> >>>>     <mailto:[hidden email]> <mailto:[hidden email]>
> >>>>     <mailto:[hidden email]>> wrote:
> >>>>
> >>>>         Dear list members,
> >>>>
> >>>>         I have a question regarding the encoding of calendar systems
> >>>>     using
> >>>>         @calendar with <date>.
> >>>>
> >>>>         Suppose we have a date given in a manuscript like this:
> >>>>
> >>>>         > Dienstag 7/19 May
> >>>>
> >>>>         This date is given in two calendar system, Julian and
> >>>>     Gregorian.
> >>>>         Would
> >>>>         it be possible to tag it like this:
> >>>>
> >>>>         <date calendar="#julian-gregorian"
> >>>>     when="1829-05-19">Dienstag 7/19
> >>>>         May</date>.
> >>>>
> >>>>         My colleagues and I are unsure if this would be a valid
> >>>>     usage of the
> >>>>         <date> element, but I think this would be a clean way to
> >>>>     mark up the
> >>>>         given date (which in my reading is a _single_ date) and note
> >>>>     the fact
> >>>>         that it is given in two different calendar systems. If at all
> >>>>         possible I
> >>>>         would like to avoid having two <date> elements or *gasp*
> >>>>     having to
> >>>>         nest
> >>>>         the <date> elements.
> >>>>
> >>>>         We also are not sure what the URIs of @calendar should be
> >>>>     pointing
> >>>>         to in
> >>>>         general. Are there defined concepts/authority file references
> >>>>         available
> >>>>         for calendar systems? I could only find examples of
> >>>>     `#gregorian`,
> >>>>         `#julian` in the TEI Guidelines.
> >>>>
> >>>>         I would be very interested in your opinions and - if
> >>>>     available -
> >>>>         examples!
> >>>>
> >>>>         All the best,
> >>>>
> >>>>         Sascha Grabsch
> >>>>
> >>>
> >>>     --
> >>>     Christian Thomas
> >>>     [hidden email]
> >>>     <mailto:[hidden email]>
> >>>     @dta_cthomas
> >>>     --
> >>>
> >>
> >
> >     --
> >     Christian Thomas
> >     [hidden email]  <mailto:[hidden email]>
> >     @dta_cthomas
> >     --
> >
>
> --
> -------------------------------------
> Humanities Computing and Media Centre
> University of Victoria
> [hidden email]
>

--
Paul Schaffner  Digital Content & Collections
University of Michigan Libraries
[hidden email] | http://www.umich.edu/~pfs/
lou
Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

lou
In reply to this post by Martin Holmes
Sorry, I hadn't fully grasped that this was a proposal for change to the current status rather than a suggestion for how to handle the use case right now. 


On Wed, 26 Aug 2020, 22:23 Martin Holmes, <[hidden email]> wrote:
Hi Lou,

I think Christian was suggesting that @calendar permit multiple
pointers, which I think makes perfect sense for scenarios like this. If
you know of any good reasons not to make such a feature request, let us
know; I think it would be a good idea myself.

Cheers,
Martin

On 2020-08-26 12:20 p.m., Lou Burnard wrote:
> I am not sure what you are trying to achieve here. The calendar
> attribute permits only a single Uri. A Uri may begin with but not
> contain the # character. So whether or not you put spaces into it your
> suggested values are syntactically invalid.
>
> On Wed, 26 Aug 2020, 19:13 Christian Thomas,
> <[hidden email]
> <mailto:[hidden email]>> wrote:
>
>
>     Hi again, just to correct my own mistakes, I don't know where all
>     the whitespaces went:
>
>     > He was born on <date calendar="#gregorian #julian" when="1732-02-22">Feb. 22, 1732 (Feb. 11, 1731/32, O.S.</date>).
>
>     Or, to take up Sascha Grabschs initial expample from the /edition
>     humboldt digital/ (cf. https://edition-humboldt.de/H0016785/12r,
>     currently tagged wrong, alas the question), "Dienstag 7/19 May"
>     should be
>
>     > <date when="1829-05-19" calendar="#gregorian #julian">Dienstag <hi rendition="#sup">7</hi>/<hi rendition="#sub">19</hi> May</date>
>
>
>     All in favour?
>
>     Best wishes
>     Christian
>
>
>     Am 26.08.2020 um 18:28 schrieb Martin Holmes:
>>     Unfortunately, @calendar currently allows only one pointer:
>>
>>     <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
>>     <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
>>
>>
>>     but I think you should be able to use multiple pointers in this
>>     case; the use-case is very straightforward. Can anyone think of a
>>     reason why not?
>>
>>     The same is not true of @datingMethod, though; because that
>>     specifies what system is being used for other attributes, and it's
>>     not possible to say which of the array of @*-custom attributes
>>     would use one dating method and which another, @datingMethod
>>     should stay as a single pointer.
>>
>>     Cheers,
>>     Martin
>>
>>     On 2020-08-26 9:20 a.m., Christian Thomas wrote:
>>>
>>>     Dear TEI list, please excuse my maybe naive question, but, taking
>>>     up the example from
>>>     https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
>>>     which says:
>>>
>>>>     He was born on <date calendar="#gregorian">Feb. 22,
>>>>     1732</date>(<date calendar="#julian" when="1732-02-22">Feb. 11,
>>>>     1731/32, O.S.</date>).
>>>
>>>     Why can't I join the two (or more) attribute values in @calendar
>>>     like this:
>>>
>>>>     He was born on <date calendar="#gregorian#julian"
>>>>     when="1732-02-22">Feb. 22, 1732(Feb. 11, 1731/32, O.S.</date>).
>>>
>>>     This would (in my reading) say that in the source both calendars
>>>     are being used, @when is the standardised Gregorian date as
>>>     specified in the TEI Guidelines, and would save us the trouble of
>>>     having to invent multiple new values for 'combined' calendars.
>>>
>>>     Please excuse me if this is layed out as impossible resp.
>>>     forbidden somewhere else (and, if so, thanks for pointing me to
>>>     that place).
>>>
>>>     Thanks, best wishes
>>>     Christian
>>>
>>>
>>>
>>>
>>>     Am 26.08.2020 um 17:19 schrieb Lou Burnard:
>>>>     The second part of your question is easy: the value of the
>>>>     @calendar attribute is a pointer to a definition of a calendar
>>>>     system. The definition might be provided by a <calendar> element
>>>>     inside a <calendarDesc> in your TEI header, or by some other
>>>>     location  entirely, like a Wikipedia entry.
>>>>
>>>>     The first part is trickier. Does the @calendar reference apply
>>>>     to the value of @when, or to the content of the element? Hands
>>>>     up if your first thought (like mine) was the former! You're
>>>>     wrong. There's a warning note at
>>>>     https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
>>>>     which says clearly
>>>>     "the calendar attribute (unlike @datingMethod defined in
>>>>     att.datable.custom
>>>>     <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>
>>>>     <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>)
>>>>     defines the calendar system of the date in the original material
>>>>     defined by the parent element, /not/ the calendar to which the
>>>>     date is normalized."
>>>>     So, if you want to indicate that the content of your date
>>>>     element is neither Julian nor Gregorian, but a combination of
>>>>     the two, you will have to define a new calendar (let's call it
>>>>     "julGreg") and reference that:
>>>>     <date calendar="#julGreg" when="1829-05-19">Dienstag 7/19
>>>>     May</date> , exactly as you propose. Note that @when is always
>>>>     understood to be normalised to the same (Gregorian) calendar. If
>>>>     you want to supply an additional normalisation to the Julian
>>>>     calendar, you'd have to add further attributes:
>>>>     <date calendar="#julGreg" when="1829-05-19"
>>>>     when-custom="1829-05-07" dating-method="#julian">Dienstag 7/19
>>>>     May</date>
>>>>
>>>>     You will of course also need to provide somewhere a calendar
>>>>     element like this:
>>>>     <calendar xml:id="julGreg"><p>A composite calendar format in
>>>>     which both Julian and Gregorian forms are provided, separated by
>>>>     a solidus, with the Julian form first.</calendar>
>>>>
>>>>
>>>>
>>>>     On Wed, 26 Aug 2020 at 12:09, Sascha Grabsch <[hidden email]
>>>>     <mailto:[hidden email]> <mailto:[hidden email]>
>>>>     <mailto:[hidden email]>> wrote:
>>>>
>>>>         Dear list members,
>>>>
>>>>         I have a question regarding the encoding of calendar systems
>>>>     using
>>>>         @calendar with <date>.
>>>>
>>>>         Suppose we have a date given in a manuscript like this:
>>>>
>>>>         > Dienstag 7/19 May
>>>>
>>>>         This date is given in two calendar system, Julian and
>>>>     Gregorian.
>>>>         Would
>>>>         it be possible to tag it like this:
>>>>
>>>>         <date calendar="#julian-gregorian"
>>>>     when="1829-05-19">Dienstag 7/19
>>>>         May</date>.
>>>>
>>>>         My colleagues and I are unsure if this would be a valid
>>>>     usage of the
>>>>         <date> element, but I think this would be a clean way to
>>>>     mark up the
>>>>         given date (which in my reading is a _single_ date) and note
>>>>     the fact
>>>>         that it is given in two different calendar systems. If at all
>>>>         possible I
>>>>         would like to avoid having two <date> elements or *gasp*
>>>>     having to
>>>>         nest
>>>>         the <date> elements.
>>>>
>>>>         We also are not sure what the URIs of @calendar should be
>>>>     pointing
>>>>         to in
>>>>         general. Are there defined concepts/authority file references
>>>>         available
>>>>         for calendar systems? I could only find examples of
>>>>     `#gregorian`,
>>>>         `#julian` in the TEI Guidelines.
>>>>
>>>>         I would be very interested in your opinions and - if
>>>>     available -
>>>>         examples!
>>>>
>>>>         All the best,
>>>>
>>>>         Sascha Grabsch
>>>>
>>>
>>>     --
>>>     Christian Thomas
>>>     [hidden email]
>>>     <mailto:[hidden email]>
>>>     @dta_cthomas
>>>     --
>>>
>>
>
>     --
>     Christian Thomas
>     [hidden email]  <mailto:[hidden email]>
>     @dta_cthomas
>     --
>

--
-------------------------------------
Humanities Computing and Media Centre
University of Victoria
[hidden email]
Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

Christian Thomas (HU Berlin)
In reply to this post by Paul Schaffner

Dear Martin, thank you:
On Wed, Aug 26, 2020, at 16:23, Martin Holmes wrote:
Hi Lou,

I think Christian was suggesting that @calendar permit multiple 
pointers, 
exactly what I meant.

Dear Paul,

Am 26.08.2020 um 22:51 schrieb Paul Schaffner:
Am I right in understanding this request to be aimed specifically
and exclusively at representations of date that more-or-less 
inextricably combine references to more than one calendar? Hence

 <date when="1829-05-19" calendar="#gregorian #julian">Thursday, 7/19 May</date>

yes, that's true and to be tagged like above, whereas this:

... whereas where the expression is separable, the need goes away?
 
 <date when="1829-05-19" calendar="#julian">Thursday, 7 May, O.S.</date> or
 <date when="1829-05-19" calendar="#gregorian">Thursday, 19 May, N.S.</date>

should be seperated. But this:

... and that the only alternative to this suggestion available within
the schema currently is to create a 'calendar' that is in effect a hybrid
between those referenced? like this?

 <date when="1829-05-19" calendar="#julian_gregorian">Thursday, 7/19 May</date>

is not the only anlternative, we could also go back to separating the date-Tags here, and we might have to, as long as date can only have one uri:
<date when="1829-05-19" calendar="#julian">Thursday, 7<date>/<date when="1829-05-19" calendar="#gregorian">19 May</date>

But Humboldt and Ehrenberg (the latter being the author of the travel journal quoted) have been on the road some 180 days, so we either tag 2x180=360 dates, but still achieve no real information gain for the user, since @when would always be the same 180 times, or we invent a new value for @calendar combining two calendars that in this combination did not exist. Both alternatives seem unattractive to me.

And: The problem multiplies many times, since we have 10+ volumes of Humboldt's own journals still ahead of us, and since he is using and noting dates in at least three different calendars, the gregorian, the julian and the french revolutionary calendar. we would like to annotate which calendar(s) is/are used in the source, and give the 'normalised' gregorian date, but not necessarily want to tag every figure in notations like "7/19 May" with an extra <date>-tag.

In sum, we would prefer the efficient solution of havin only one <date>-element, but with multiple @calendar-values whereever applicable in the way Paul illustrated in his first example. Can we agree that @calendar should, as an option, be allowed to hold more than one calendar as its value?

Thank you all for your input, looking forward for more!
Best, Christian

pfs
 
 

On Wed, Aug 26, 2020, at 16:23, Martin Holmes wrote:
Hi Lou,

I think Christian was suggesting that @calendar permit multiple 
pointers, which I think makes perfect sense for scenarios like this. If 
you know of any good reasons not to make such a feature request, let us 
know; I think it would be a good idea myself.

Cheers,
Martin

On 2020-08-26 12:20 p.m., Lou Burnard wrote:
I am not sure what you are trying to achieve here. The calendar 
attribute permits only a single Uri. A Uri may begin with but not 
contain the # character. So whether or not you put spaces into it your 
suggested values are syntactically invalid.

On Wed, 26 Aug 2020, 19:13 Christian Thomas, 
<[hidden email] 
[hidden email]> wrote:


    Hi again, just to correct my own mistakes, I don't know where all
    the whitespaces went:

    > He was born on <date calendar="#gregorian #julian" when="1732-02-22">Feb. 22, 1732 (Feb. 11, 1731/32, O.S.</date>).

    Or, to take up Sascha Grabschs initial expample from the /edition
    humboldt digital/ (cf. https://edition-humboldt.de/H0016785/12r,
    currently tagged wrong, alas the question), "Dienstag 7/19 May"
    should be

    > <date when="1829-05-19" calendar="#gregorian #julian">Dienstag <hi rendition="#sup">7</hi>/<hi rendition="#sub">19</hi> May</date>


    All in favour?

    Best wishes
    Christian


    Am 26.08.2020 um 18:28 schrieb Martin Holmes:
    Unfortunately, @calendar currently allows only one pointer:

    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>


    but I think you should be able to use multiple pointers in this
    case; the use-case is very straightforward. Can anyone think of a
    reason why not?

    The same is not true of @datingMethod, though; because that
    specifies what system is being used for other attributes, and it's
    not possible to say which of the array of @*-custom attributes
    would use one dating method and which another, @datingMethod
    should stay as a single pointer.

    Cheers,
    Martin

    On 2020-08-26 9:20 a.m., Christian Thomas wrote:
    Dear TEI list, please excuse my maybe naive question, but, taking
    up the example from
    https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
    which says:

    He was born on <date calendar="#gregorian">Feb. 22,
    1732</date>(<date calendar="#julian" when="1732-02-22">Feb. 11,
    1731/32, O.S.</date>).
    Why can't I join the two (or more) attribute values in @calendar
    like this:

    He was born on <date calendar="#gregorian#julian"
    when="1732-02-22">Feb. 22, 1732(Feb. 11, 1731/32, O.S.</date>).
    This would (in my reading) say that in the source both calendars
    are being used, @when is the standardised Gregorian date as
    specified in the TEI Guidelines, and would save us the trouble of
    having to invent multiple new values for 'combined' calendars.

    Please excuse me if this is layed out as impossible resp.
    forbidden somewhere else (and, if so, thanks for pointing me to
    that place).

    Thanks, best wishes
    Christian




    Am 26.08.2020 um 17:19 schrieb Lou Burnard:
    The second part of your question is easy: the value of the
    @calendar attribute is a pointer to a definition of a calendar
    system. The definition might be provided by a <calendar> element
    inside a <calendarDesc> in your TEI header, or by some other
    location  entirely, like a Wikipedia entry.

    The first part is trickier. Does the @calendar reference apply
    to the value of @when, or to the content of the element? Hands
    up if your first thought (like mine) was the former! You're
    wrong. There's a warning note at
    https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
    which says clearly
    "the calendar attribute (unlike @datingMethod defined in
    att.datable.custom
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>)
    defines the calendar system of the date in the original material
    defined by the parent element, /not/ the calendar to which the
    date is normalized."
    So, if you want to indicate that the content of your date
    element is neither Julian nor Gregorian, but a combination of
    the two, you will have to define a new calendar (let's call it
    "julGreg") and reference that:
    <date calendar="#julGreg" when="1829-05-19">Dienstag 7/19
    May</date> , exactly as you propose. Note that @when is always
    understood to be normalised to the same (Gregorian) calendar. If
    you want to supply an additional normalisation to the Julian
    calendar, you'd have to add further attributes:
    <date calendar="#julGreg" when="1829-05-19"
    when-custom="1829-05-07" dating-method="#julian">Dienstag 7/19
    May</date>

    You will of course also need to provide somewhere a calendar
    element like this:
    <calendar xml:id="julGreg"><p>A composite calendar format in
    which both Julian and Gregorian forms are provided, separated by
    a solidus, with the Julian form first.</calendar>



    On Wed, 26 Aug 2020 at 12:09, Sascha Grabsch <[hidden email]
    [hidden email] [hidden email]
    [hidden email]> wrote:

        Dear list members,

        I have a question regarding the encoding of calendar systems
    using
        @calendar with <date>.

        Suppose we have a date given in a manuscript like this:

        > Dienstag 7/19 May

        This date is given in two calendar system, Julian and
    Gregorian.
        Would
        it be possible to tag it like this:

        <date calendar="#julian-gregorian"
    when="1829-05-19">Dienstag 7/19
        May</date>.

        My colleagues and I are unsure if this would be a valid
    usage of the
        <date> element, but I think this would be a clean way to
    mark up the
        given date (which in my reading is a _single_ date) and note
    the fact
        that it is given in two different calendar systems. If at all
        possible I
        would like to avoid having two <date> elements or *gasp*
    having to
        nest
        the <date> elements.

        We also are not sure what the URIs of @calendar should be
    pointing
        to in
        general. Are there defined concepts/authority file references
        available
        for calendar systems? I could only find examples of
    `#gregorian`,
        `#julian` in the TEI Guidelines.

        I would be very interested in your opinions and - if
    available -
        examples!

        All the best,

        Sascha Grabsch

    -- 
    Christian Thomas
    [hidden email]
    [hidden email]
    @dta_cthomas
    -- 

    -- 
    Christian Thomas
    [hidden email]  [hidden email]
    @dta_cthomas
    --

-- 
-------------------------------------
Humanities Computing and Media Centre
University of Victoria
[hidden email]


Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

Kurz, Stephan

Dear Christian, dear list,

 

I could not think of any counter arguments apart from the solution not being immediately human-readable (only one normalized @when combined with two @calendar pointers).

Of course, one could explain this e.g. in calendarDesc; in the use case at hand with something like “Humboldt notes dates in both Julian and Gregorian variant, here @when only contains the normalized Gregorian date”.  

I guess adding the Julian date in @when-custom as I suggested yesterday would not provide additional information necessary and useful for a reader of today, even if it would add symmetry in having two calendars and two whens.

 

Best, Stephan

 

From: TEI (Text Encoding Initiative) public discussion list <[hidden email]> On Behalf Of Christian Thomas
Sent: Thursday, August 27, 2020 1:16 AM
To: [hidden email]
Subject: Re: Multiple values for @calendar in the same <date> element

 


Dear Martin, thank you:

On Wed, Aug 26, 2020, at 16:23, Martin Holmes wrote:
Hi Lou,
 
I think Christian was suggesting that @calendar permit multiple 
pointers, 

exactly what I meant.

Dear Paul,

Am 26.08.2020 um 22:51 schrieb Paul Schaffner:

Am I right in understanding this request to be aimed specifically
and exclusively at representations of date that more-or-less 
inextricably combine references to more than one calendar? Hence
 
 <date when="1829-05-19" calendar="#gregorian #julian">Thursday, 7/19 May</date>


yes, that's true and to be tagged like above, whereas this:


... whereas where the expression is separable, the need goes away?
 
 <date when="1829-05-19" calendar="#julian">Thursday, 7 May, O.S.</date> or
 <date when="1829-05-19" calendar="#gregorian">Thursday, 19 May, N.S.</date>


should be seperated. But this:


... and that the only alternative to this suggestion available within
the schema currently is to create a 'calendar' that is in effect a hybrid
between those referenced? like this?
 
 <date when="1829-05-19" calendar="#julian_gregorian">Thursday, 7/19 May</date>


is not the only anlternative, we could also go back to separating the date-Tags here, and we might have to, as long as date can only have one uri:

<date when="1829-05-19" calendar="#julian">Thursday, 7<date>/<date when="1829-05-19" calendar="#gregorian">19 May</date>


But Humboldt and Ehrenberg (the latter being the author of the travel journal quoted) have been on the road some 180 days, so we either tag 2x180=360 dates, but still achieve no real information gain for the user, since @when would always be the same 180 times, or we invent a new value for @calendar combining two calendars that in this combination did not exist. Both alternatives seem unattractive to me.

And: The problem multiplies many times, since we have 10+ volumes of Humboldt's own journals still ahead of us, and since he is using and noting dates in at least three different calendars, the gregorian, the julian and the french revolutionary calendar. we would like to annotate which calendar(s) is/are used in the source, and give the 'normalised' gregorian date, but not necessarily want to tag every figure in notations like "7/19 May" with an extra <date>-tag.

In sum, we would prefer the efficient solution of havin only one <date>-element, but with multiple @calendar-values whereever applicable in the way Paul illustrated in his first example. Can we agree that @calendar should, as an option, be allowed to hold more than one calendar as its value?

Thank you all for your input, looking forward for more!
Best, Christian


 
pfs
 
 
 
On Wed, Aug 26, 2020, at 16:23, Martin Holmes wrote:
Hi Lou,
 
I think Christian was suggesting that @calendar permit multiple 
pointers, which I think makes perfect sense for scenarios like this. If 
you know of any good reasons not to make such a feature request, let us 
know; I think it would be a good idea myself.
 
Cheers,
Martin
 
On 2020-08-26 12:20 p.m., Lou Burnard wrote:
I am not sure what you are trying to achieve here. The calendar 
attribute permits only a single Uri. A Uri may begin with but not 
contain the # character. So whether or not you put spaces into it your 
suggested values are syntactically invalid.
 
On Wed, 26 Aug 2020, 19:13 Christian Thomas, 
<[hidden email] 
[hidden email]> wrote:
 
 
    Hi again, just to correct my own mistakes, I don't know where all
    the whitespaces went:
 
    > He was born on <date calendar="#gregorian #julian" when="1732-02-22">Feb. 22, 1732 (Feb. 11, 1731/32, O.S.</date>).
 
    Or, to take up Sascha Grabschs initial expample from the /edition
    humboldt digital/ (cf. https://edition-humboldt.de/H0016785/12r,
    currently tagged wrong, alas the question), "Dienstag 7/19 May"
    should be
 
    > <date when="1829-05-19" calendar="#gregorian #julian">Dienstag <hi rendition="#sup">7</hi>/<hi rendition="#sub">19</hi> May</date>
 
 
    All in favour?
 
    Best wishes
    Christian
 
 
    Am 26.08.2020 um 18:28 schrieb Martin Holmes:
    Unfortunately, @calendar currently allows only one pointer:
 
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
 
 
    but I think you should be able to use multiple pointers in this
    case; the use-case is very straightforward. Can anyone think of a
    reason why not?
 
    The same is not true of @datingMethod, though; because that
    specifies what system is being used for other attributes, and it's
    not possible to say which of the array of @*-custom attributes
    would use one dating method and which another, @datingMethod
    should stay as a single pointer.
 
    Cheers,
    Martin
 
    On 2020-08-26 9:20 a.m., Christian Thomas wrote:
    Dear TEI list, please excuse my maybe naive question, but, taking
    up the example from
    https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
    which says:
 
    He was born on <date calendar="#gregorian">Feb. 22,
    1732</date>(<date calendar="#julian" when="1732-02-22">Feb. 11,
    1731/32, O.S.</date>).
    Why can't I join the two (or more) attribute values in @calendar
    like this:
 
    He was born on <date calendar="#gregorian#julian"
    when="1732-02-22">Feb. 22, 1732(Feb. 11, 1731/32, O.S.</date>).
    This would (in my reading) say that in the source both calendars
    are being used, @when is the standardised Gregorian date as
    specified in the TEI Guidelines, and would save us the trouble of
    having to invent multiple new values for 'combined' calendars.
 
    Please excuse me if this is layed out as impossible resp.
    forbidden somewhere else (and, if so, thanks for pointing me to
    that place).
 
    Thanks, best wishes
    Christian
 
 
 
 
    Am 26.08.2020 um 17:19 schrieb Lou Burnard:
    The second part of your question is easy: the value of the
    @calendar attribute is a pointer to a definition of a calendar
    system. The definition might be provided by a <calendar> element
    inside a <calendarDesc> in your TEI header, or by some other
    location  entirely, like a Wikipedia entry.
 
    The first part is trickier. Does the @calendar reference apply
    to the value of @when, or to the content of the element? Hands
    up if your first thought (like mine) was the former! You're
    wrong. There's a warning note at
    https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
    which says clearly
    "the calendar attribute (unlike @datingMethod defined in
    att.datable.custom
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>)
    defines the calendar system of the date in the original material
    defined by the parent element, /not/ the calendar to which the
    date is normalized."
    So, if you want to indicate that the content of your date
    element is neither Julian nor Gregorian, but a combination of
    the two, you will have to define a new calendar (let's call it
    "julGreg") and reference that:
    <date calendar="#julGreg" when="1829-05-19">Dienstag 7/19
    May</date> , exactly as you propose. Note that @when is always
    understood to be normalised to the same (Gregorian) calendar. If
    you want to supply an additional normalisation to the Julian
    calendar, you'd have to add further attributes:
    <date calendar="#julGreg" when="1829-05-19"
    when-custom="1829-05-07" dating-method="#julian">Dienstag 7/19
    May</date>
 
    You will of course also need to provide somewhere a calendar
    element like this:
    <calendar xml:id="julGreg"><p>A composite calendar format in
    which both Julian and Gregorian forms are provided, separated by
    a solidus, with the Julian form first.</calendar>
 
 
 
    On Wed, 26 Aug 2020 at 12:09, Sascha Grabsch <[hidden email]
    [hidden email] [hidden email]
    [hidden email]> wrote:
 
        Dear list members,
 
        I have a question regarding the encoding of calendar systems
    using
        @calendar with <date>.
 
        Suppose we have a date given in a manuscript like this:
 
        > Dienstag 7/19 May
 
        This date is given in two calendar system, Julian and
    Gregorian.
        Would
        it be possible to tag it like this:
 
        <date calendar="#julian-gregorian"
    when="1829-05-19">Dienstag 7/19
        May</date>.
 
        My colleagues and I are unsure if this would be a valid
    usage of the
        <date> element, but I think this would be a clean way to
    mark up the
        given date (which in my reading is a _single_ date) and note
    the fact
        that it is given in two different calendar systems. If at all
        possible I
        would like to avoid having two <date> elements or *gasp*
    having to
        nest
        the <date> elements.
 
        We also are not sure what the URIs of @calendar should be
    pointing
        to in
        general. Are there defined concepts/authority file references
        available
        for calendar systems? I could only find examples of
    `#gregorian`,
        `#julian` in the TEI Guidelines.
 
        I would be very interested in your opinions and - if
    available -
        examples!
 
        All the best,
 
        Sascha Grabsch
 
    -- 
    Christian Thomas
    [hidden email]
    [hidden email]
    @dta_cthomas
    -- 
 
    -- 
    Christian Thomas
    [hidden email]  [hidden email]
    @dta_cthomas
    --
 
-- 
-------------------------------------
Humanities Computing and Media Centre
University of Victoria
[hidden email]
 

 

Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

Daniele Marotta-2

Dear Stephan, dear list,

I'm afraid it would not be sufficient the @when-custom attribute if "he is using and noting dates in at least three different calendars, the gregorian, the julian and the french revolutionary calendar". I guess there could be cases where they have all three together.

Best, Daniele

Il 27/08/2020 10:33, Kurz, Stephan ha scritto:

Dear Christian, dear list,

 

I could not think of any counter arguments apart from the solution not being immediately human-readable (only one normalized @when combined with two @calendar pointers).

Of course, one could explain this e.g. in calendarDesc; in the use case at hand with something like “Humboldt notes dates in both Julian and Gregorian variant, here @when only contains the normalized Gregorian date”.  

I guess adding the Julian date in @when-custom as I suggested yesterday would not provide additional information necessary and useful for a reader of today, even if it would add symmetry in having two calendars and two whens.

 

Best, Stephan

 

From: TEI (Text Encoding Initiative) public discussion list [hidden email] On Behalf Of Christian Thomas
Sent: Thursday, August 27, 2020 1:16 AM
To: [hidden email]
Subject: Re: Multiple values for @calendar in the same <date> element

 


Dear Martin, thank you:

On Wed, Aug 26, 2020, at 16:23, Martin Holmes wrote:
Hi Lou,
 
I think Christian was suggesting that @calendar permit multiple 
pointers, 

exactly what I meant.

Dear Paul,

Am 26.08.2020 um 22:51 schrieb Paul Schaffner:

Am I right in understanding this request to be aimed specifically
and exclusively at representations of date that more-or-less 
inextricably combine references to more than one calendar? Hence
 
 <date when="1829-05-19" calendar="#gregorian #julian">Thursday, 7/19 May</date>


yes, that's true and to be tagged like above, whereas this:


... whereas where the expression is separable, the need goes away?
 
 <date when="1829-05-19" calendar="#julian">Thursday, 7 May, O.S.</date> or
 <date when="1829-05-19" calendar="#gregorian">Thursday, 19 May, N.S.</date>


should be seperated. But this:


... and that the only alternative to this suggestion available within
the schema currently is to create a 'calendar' that is in effect a hybrid
between those referenced? like this?
 
 <date when="1829-05-19" calendar="#julian_gregorian">Thursday, 7/19 May</date>


is not the only anlternative, we could also go back to separating the date-Tags here, and we might have to, as long as date can only have one uri:

<date when="1829-05-19" calendar="#julian">Thursday, 7<date>/<date when="1829-05-19" calendar="#gregorian">19 May</date>


But Humboldt and Ehrenberg (the latter being the author of the travel journal quoted) have been on the road some 180 days, so we either tag 2x180=360 dates, but still achieve no real information gain for the user, since @when would always be the same 180 times, or we invent a new value for @calendar combining two calendars that in this combination did not exist. Both alternatives seem unattractive to me.

And: The problem multiplies many times, since we have 10+ volumes of Humboldt's own journals still ahead of us, and since he is using and noting dates in at least three different calendars, the gregorian, the julian and the french revolutionary calendar. we would like to annotate which calendar(s) is/are used in the source, and give the 'normalised' gregorian date, but not necessarily want to tag every figure in notations like "7/19 May" with an extra <date>-tag.

In sum, we would prefer the efficient solution of havin only one <date>-element, but with multiple @calendar-values whereever applicable in the way Paul illustrated in his first example. Can we agree that @calendar should, as an option, be allowed to hold more than one calendar as its value?

Thank you all for your input, looking forward for more!
Best, Christian


 
pfs
 
 
 
On Wed, Aug 26, 2020, at 16:23, Martin Holmes wrote:
Hi Lou,
 
I think Christian was suggesting that @calendar permit multiple 
pointers, which I think makes perfect sense for scenarios like this. If 
you know of any good reasons not to make such a feature request, let us 
know; I think it would be a good idea myself.
 
Cheers,
Martin
 
On 2020-08-26 12:20 p.m., Lou Burnard wrote:
I am not sure what you are trying to achieve here. The calendar 
attribute permits only a single Uri. A Uri may begin with but not 
contain the # character. So whether or not you put spaces into it your 
suggested values are syntactically invalid.
 
On Wed, 26 Aug 2020, 19:13 Christian Thomas, 
<[hidden email] 
[hidden email]> wrote:
 
 
    Hi again, just to correct my own mistakes, I don't know where all
    the whitespaces went:
 
    > He was born on <date calendar="#gregorian #julian" when="1732-02-22">Feb. 22, 1732 (Feb. 11, 1731/32, O.S.</date>).
 
    Or, to take up Sascha Grabschs initial expample from the /edition
    humboldt digital/ (cf. https://edition-humboldt.de/H0016785/12r,
    currently tagged wrong, alas the question), "Dienstag 7/19 May"
    should be
 
    > <date when="1829-05-19" calendar="#gregorian #julian">Dienstag <hi rendition="#sup">7</hi>/<hi rendition="#sub">19</hi> May</date>
 
 
    All in favour?
 
    Best wishes
    Christian
 
 
    Am 26.08.2020 um 18:28 schrieb Martin Holmes:
    Unfortunately, @calendar currently allows only one pointer:
 
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
 
 
    but I think you should be able to use multiple pointers in this
    case; the use-case is very straightforward. Can anyone think of a
    reason why not?
 
    The same is not true of @datingMethod, though; because that
    specifies what system is being used for other attributes, and it's
    not possible to say which of the array of @*-custom attributes
    would use one dating method and which another, @datingMethod
    should stay as a single pointer.
 
    Cheers,
    Martin
 
    On 2020-08-26 9:20 a.m., Christian Thomas wrote:
    Dear TEI list, please excuse my maybe naive question, but, taking
    up the example from
    https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
    which says:
 
    He was born on <date calendar="#gregorian">Feb. 22,
    1732</date>(<date calendar="#julian" when="1732-02-22">Feb. 11,
    1731/32, O.S.</date>).
    Why can't I join the two (or more) attribute values in @calendar
    like this:
 
    He was born on <date calendar="#gregorian#julian"
    when="1732-02-22">Feb. 22, 1732(Feb. 11, 1731/32, O.S.</date>).
    This would (in my reading) say that in the source both calendars
    are being used, @when is the standardised Gregorian date as
    specified in the TEI Guidelines, and would save us the trouble of
    having to invent multiple new values for 'combined' calendars.
 
    Please excuse me if this is layed out as impossible resp.
    forbidden somewhere else (and, if so, thanks for pointing me to
    that place).
 
    Thanks, best wishes
    Christian
 
 
 
 
    Am 26.08.2020 um 17:19 schrieb Lou Burnard:
    The second part of your question is easy: the value of the
    @calendar attribute is a pointer to a definition of a calendar
    system. The definition might be provided by a <calendar> element
    inside a <calendarDesc> in your TEI header, or by some other
    location  entirely, like a Wikipedia entry.
 
    The first part is trickier. Does the @calendar reference apply
    to the value of @when, or to the content of the element? Hands
    up if your first thought (like mine) was the former! You're
    wrong. There's a warning note at
    https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
    which says clearly
    "the calendar attribute (unlike @datingMethod defined in
    att.datable.custom
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>)
    defines the calendar system of the date in the original material
    defined by the parent element, /not/ the calendar to which the
    date is normalized."
    So, if you want to indicate that the content of your date
    element is neither Julian nor Gregorian, but a combination of
    the two, you will have to define a new calendar (let's call it
    "julGreg") and reference that:
    <date calendar="#julGreg" when="1829-05-19">Dienstag 7/19
    May</date> , exactly as you propose. Note that @when is always
    understood to be normalised to the same (Gregorian) calendar. If
    you want to supply an additional normalisation to the Julian
    calendar, you'd have to add further attributes:
    <date calendar="#julGreg" when="1829-05-19"
    when-custom="1829-05-07" dating-method="#julian">Dienstag 7/19
    May</date>
 
    You will of course also need to provide somewhere a calendar
    element like this:
    <calendar xml:id="julGreg"><p>A composite calendar format in
    which both Julian and Gregorian forms are provided, separated by
    a solidus, with the Julian form first.</calendar>
 
 
 
    On Wed, 26 Aug 2020 at 12:09, Sascha Grabsch <[hidden email]
    [hidden email] [hidden email]
    [hidden email]> wrote:
 
        Dear list members,
 
        I have a question regarding the encoding of calendar systems
    using
        @calendar with <date>.
 
        Suppose we have a date given in a manuscript like this:
 
        > Dienstag 7/19 May
 
        This date is given in two calendar system, Julian and
    Gregorian.
        Would
        it be possible to tag it like this:
 
        <date calendar="#julian-gregorian"
    when="1829-05-19">Dienstag 7/19
        May</date>.
 
        My colleagues and I are unsure if this would be a valid
    usage of the
        <date> element, but I think this would be a clean way to
    mark up the
        given date (which in my reading is a _single_ date) and note
    the fact
        that it is given in two different calendar systems. If at all
        possible I
        would like to avoid having two <date> elements or *gasp*
    having to
        nest
        the <date> elements.
 
        We also are not sure what the URIs of @calendar should be
    pointing
        to in
        general. Are there defined concepts/authority file references
        available
        for calendar systems? I could only find examples of
    `#gregorian`,
        `#julian` in the TEI Guidelines.
 
        I would be very interested in your opinions and - if
    available -
        examples!
 
        All the best,
 
        Sascha Grabsch
 
    -- 
    Christian Thomas
    [hidden email]
    [hidden email]
    @dta_cthomas
    -- 
 
    -- 
    Christian Thomas
    [hidden email]  [hidden email]
    @dta_cthomas
    --
 
-- 
-------------------------------------
Humanities Computing and Media Centre
University of Victoria
[hidden email]
 

 

lou
Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

lou
In reply to this post by Christian Thomas (HU Berlin)
Dear Christian

Thank you for arguing your case so eloquently. I hope you'll forgive me if I argue for a different point of view!
Firstly though, please note that I entirely agree with you that the answer is not to provide multiple <date> elements, for all sorts of reasons, many of which you have already pointed out.
But I don't think allowing @calendar to take multiple values is  the answer either, for the following reasons:
1. It is unnecessary. The proposed solution of defining a different <calendar> for dates in which the content is expressed as a combination works fine without any need to change anything.
2. It's time consuming. Changing the datatype of a predefined attribute means a change to the TEI schema, which has to percolate through the Council's change mechanism, and won't be available generally until there is a new TEI release. [though of course you can make the change locally in your own ODD, if you're willing to support a non-conformant schema for yourself]
3. It introduces ambiguity. If it were implemented, the existing mechanism would not cease to work, so there would then be two different ways of doing the same thing. Of course that happens all over the place in the TEI, but that doesn't make it a good thing.
4. It's actually not true! In general, when you provide multiple values for a pointer, the implication is that both values apply. For example, if I say <persName ref="#foo #bar"/> I mean that the person named here is documented in both of the places identified by #foo and #bar. But in your case, the content of the date element does not respect the two calendars indicated independently, but in conjunction. Another way of thinking of this might be to hypothesize some process which is designed to work with Julian calendar values only.  Faced with a @calendar value like that you propose, such software would wrongly assume it could meaningfully process the content of the date. But this is actually in a combination of Julian and Gregorian, not Julian alone!
All in all, I think for your project the right answer is to define explicit <calendar> elements for the cases which actually exist in your data. The effort of defining two or three additional elements in your header is not so great, surely?

best wishes

Lou


On Thu, 27 Aug 2020 at 00:16, Christian Thomas <[hidden email]> wrote:

Dear Martin, thank you:
On Wed, Aug 26, 2020, at 16:23, Martin Holmes wrote:
Hi Lou,

I think Christian was suggesting that @calendar permit multiple 
pointers, 
exactly what I meant.

Dear Paul,

Am 26.08.2020 um 22:51 schrieb Paul Schaffner:
Am I right in understanding this request to be aimed specifically
and exclusively at representations of date that more-or-less 
inextricably combine references to more than one calendar? Hence

 <date when="1829-05-19" calendar="#gregorian #julian">Thursday, 7/19 May</date>

yes, that's true and to be tagged like above, whereas this:

... whereas where the expression is separable, the need goes away?
 
 <date when="1829-05-19" calendar="#julian">Thursday, 7 May, O.S.</date> or
 <date when="1829-05-19" calendar="#gregorian">Thursday, 19 May, N.S.</date>

should be seperated. But this:

... and that the only alternative to this suggestion available within
the schema currently is to create a 'calendar' that is in effect a hybrid
between those referenced? like this?

 <date when="1829-05-19" calendar="#julian_gregorian">Thursday, 7/19 May</date>

is not the only anlternative, we could also go back to separating the date-Tags here, and we might have to, as long as date can only have one uri:
<date when="1829-05-19" calendar="#julian">Thursday, 7<date>/<date when="1829-05-19" calendar="#gregorian">19 May</date>

But Humboldt and Ehrenberg (the latter being the author of the travel journal quoted) have been on the road some 180 days, so we either tag 2x180=360 dates, but still achieve no real information gain for the user, since @when would always be the same 180 times, or we invent a new value for @calendar combining two calendars that in this combination did not exist. Both alternatives seem unattractive to me.

And: The problem multiplies many times, since we have 10+ volumes of Humboldt's own journals still ahead of us, and since he is using and noting dates in at least three different calendars, the gregorian, the julian and the french revolutionary calendar. we would like to annotate which calendar(s) is/are used in the source, and give the 'normalised' gregorian date, but not necessarily want to tag every figure in notations like "7/19 May" with an extra <date>-tag.

In sum, we would prefer the efficient solution of havin only one <date>-element, but with multiple @calendar-values whereever applicable in the way Paul illustrated in his first example. Can we agree that @calendar should, as an option, be allowed to hold more than one calendar as its value?

Thank you all for your input, looking forward for more!
Best, Christian

pfs
 
 

On Wed, Aug 26, 2020, at 16:23, Martin Holmes wrote:
Hi Lou,

I think Christian was suggesting that @calendar permit multiple 
pointers, which I think makes perfect sense for scenarios like this. If 
you know of any good reasons not to make such a feature request, let us 
know; I think it would be a good idea myself.

Cheers,
Martin

On 2020-08-26 12:20 p.m., Lou Burnard wrote:
I am not sure what you are trying to achieve here. The calendar 
attribute permits only a single Uri. A Uri may begin with but not 
contain the # character. So whether or not you put spaces into it your 
suggested values are syntactically invalid.

On Wed, 26 Aug 2020, 19:13 Christian Thomas, 
<[hidden email] 
[hidden email]> wrote:


    Hi again, just to correct my own mistakes, I don't know where all
    the whitespaces went:

    > He was born on <date calendar="#gregorian #julian" when="1732-02-22">Feb. 22, 1732 (Feb. 11, 1731/32, O.S.</date>).

    Or, to take up Sascha Grabschs initial expample from the /edition
    humboldt digital/ (cf. https://edition-humboldt.de/H0016785/12r,
    currently tagged wrong, alas the question), "Dienstag 7/19 May"
    should be

    > <date when="1829-05-19" calendar="#gregorian #julian">Dienstag <hi rendition="#sup">7</hi>/<hi rendition="#sub">19</hi> May</date>


    All in favour?

    Best wishes
    Christian


    Am 26.08.2020 um 18:28 schrieb Martin Holmes:
    Unfortunately, @calendar currently allows only one pointer:

    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>


    but I think you should be able to use multiple pointers in this
    case; the use-case is very straightforward. Can anyone think of a
    reason why not?

    The same is not true of @datingMethod, though; because that
    specifies what system is being used for other attributes, and it's
    not possible to say which of the array of @*-custom attributes
    would use one dating method and which another, @datingMethod
    should stay as a single pointer.

    Cheers,
    Martin

    On 2020-08-26 9:20 a.m., Christian Thomas wrote:
    Dear TEI list, please excuse my maybe naive question, but, taking
    up the example from
    https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
    which says:

    He was born on <date calendar="#gregorian">Feb. 22,
    1732</date>(<date calendar="#julian" when="1732-02-22">Feb. 11,
    1731/32, O.S.</date>).
    Why can't I join the two (or more) attribute values in @calendar
    like this:

    He was born on <date calendar="#gregorian#julian"
    when="1732-02-22">Feb. 22, 1732(Feb. 11, 1731/32, O.S.</date>).
    This would (in my reading) say that in the source both calendars
    are being used, @when is the standardised Gregorian date as
    specified in the TEI Guidelines, and would save us the trouble of
    having to invent multiple new values for 'combined' calendars.

    Please excuse me if this is layed out as impossible resp.
    forbidden somewhere else (and, if so, thanks for pointing me to
    that place).

    Thanks, best wishes
    Christian




    Am 26.08.2020 um 17:19 schrieb Lou Burnard:
    The second part of your question is easy: the value of the
    @calendar attribute is a pointer to a definition of a calendar
    system. The definition might be provided by a <calendar> element
    inside a <calendarDesc> in your TEI header, or by some other
    location  entirely, like a Wikipedia entry.

    The first part is trickier. Does the @calendar reference apply
    to the value of @when, or to the content of the element? Hands
    up if your first thought (like mine) was the former! You're
    wrong. There's a warning note at
    https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
    which says clearly
    "the calendar attribute (unlike @datingMethod defined in
    att.datable.custom
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>)
    defines the calendar system of the date in the original material
    defined by the parent element, /not/ the calendar to which the
    date is normalized."
    So, if you want to indicate that the content of your date
    element is neither Julian nor Gregorian, but a combination of
    the two, you will have to define a new calendar (let's call it
    "julGreg") and reference that:
    <date calendar="#julGreg" when="1829-05-19">Dienstag 7/19
    May</date> , exactly as you propose. Note that @when is always
    understood to be normalised to the same (Gregorian) calendar. If
    you want to supply an additional normalisation to the Julian
    calendar, you'd have to add further attributes:
    <date calendar="#julGreg" when="1829-05-19"
    when-custom="1829-05-07" dating-method="#julian">Dienstag 7/19
    May</date>

    You will of course also need to provide somewhere a calendar
    element like this:
    <calendar xml:id="julGreg"><p>A composite calendar format in
    which both Julian and Gregorian forms are provided, separated by
    a solidus, with the Julian form first.</calendar>



    On Wed, 26 Aug 2020 at 12:09, Sascha Grabsch <[hidden email]
    [hidden email] [hidden email]
    [hidden email]> wrote:

        Dear list members,

        I have a question regarding the encoding of calendar systems
    using
        @calendar with <date>.

        Suppose we have a date given in a manuscript like this:

        > Dienstag 7/19 May

        This date is given in two calendar system, Julian and
    Gregorian.
        Would
        it be possible to tag it like this:

        <date calendar="#julian-gregorian"
    when="1829-05-19">Dienstag 7/19
        May</date>.

        My colleagues and I are unsure if this would be a valid
    usage of the
        <date> element, but I think this would be a clean way to
    mark up the
        given date (which in my reading is a _single_ date) and note
    the fact
        that it is given in two different calendar systems. If at all
        possible I
        would like to avoid having two <date> elements or *gasp*
    having to
        nest
        the <date> elements.

        We also are not sure what the URIs of @calendar should be
    pointing
        to in
        general. Are there defined concepts/authority file references
        available
        for calendar systems? I could only find examples of
    `#gregorian`,
        `#julian` in the TEI Guidelines.

        I would be very interested in your opinions and - if
    available -
        examples!

        All the best,

        Sascha Grabsch

    -- 
    Christian Thomas
    [hidden email]
    [hidden email]
    @dta_cthomas
    -- 

    -- 
    Christian Thomas
    [hidden email]  [hidden email]
    @dta_cthomas
    --

-- 
-------------------------------------
Humanities Computing and Media Centre
University of Victoria
[hidden email]


Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

Thibault Clérice
In reply to this post by Daniele Marotta-2

Dear all,

This is probably my first message here, but I have not seen the following proposal:

He was born on <date calendar="#gregorian">Feb. 22, 1732(<date calendar="#julian">Feb. 11, 1731/32, O.S.</date>).</date>
 <date when="1829-05-19" calendar="#gregorian>Thursday, 7<date calendar= #julian">/19</date> May</date>

This, or another tag to be fair, would allow to no duplicate the @when while keeping the information about the different calendar. The issue with both pointer is that you do not define which segment is a representation of which calendar system. Here, with a tree-like solution (rather than sequential nodes that would meet many exceptions), we could actually clear things up ?

If this had been brought up, I am sorry, I did not find it in the thread...

Best,

Thibault


On 27/08/2020 10:43, Daniele Marotta wrote:

Dear Stephan, dear list,

I'm afraid it would not be sufficient the @when-custom attribute if "he is using and noting dates in at least three different calendars, the gregorian, the julian and the french revolutionary calendar". I guess there could be cases where they have all three together.

Best, Daniele

Il 27/08/2020 10:33, Kurz, Stephan ha scritto:

Dear Christian, dear list,

 

I could not think of any counter arguments apart from the solution not being immediately human-readable (only one normalized @when combined with two @calendar pointers).

Of course, one could explain this e.g. in calendarDesc; in the use case at hand with something like “Humboldt notes dates in both Julian and Gregorian variant, here @when only contains the normalized Gregorian date”.  

I guess adding the Julian date in @when-custom as I suggested yesterday would not provide additional information necessary and useful for a reader of today, even if it would add symmetry in having two calendars and two whens.

 

Best, Stephan

 

From: TEI (Text Encoding Initiative) public discussion list [hidden email] On Behalf Of Christian Thomas
Sent: Thursday, August 27, 2020 1:16 AM
To: [hidden email]
Subject: Re: Multiple values for @calendar in the same <date> element

 


Dear Martin, thank you:

On Wed, Aug 26, 2020, at 16:23, Martin Holmes wrote:
Hi Lou,
 
I think Christian was suggesting that @calendar permit multiple 
pointers, 

exactly what I meant.

Dear Paul,

Am 26.08.2020 um 22:51 schrieb Paul Schaffner:

Am I right in understanding this request to be aimed specifically
and exclusively at representations of date that more-or-less 
inextricably combine references to more than one calendar? Hence
 
 <date when="1829-05-19" calendar="#gregorian #julian">Thursday, 7/19 May</date>


yes, that's true and to be tagged like above, whereas this:


... whereas where the expression is separable, the need goes away?
 
 <date when="1829-05-19" calendar="#julian">Thursday, 7 May, O.S.</date> or
 <date when="1829-05-19" calendar="#gregorian">Thursday, 19 May, N.S.</date>


should be seperated. But this:


... and that the only alternative to this suggestion available within
the schema currently is to create a 'calendar' that is in effect a hybrid
between those referenced? like this?
 
 <date when="1829-05-19" calendar="#julian_gregorian">Thursday, 7/19 May</date>


is not the only anlternative, we could also go back to separating the date-Tags here, and we might have to, as long as date can only have one uri:

<date when="1829-05-19" calendar="#julian">Thursday, 7<date>/<date when="1829-05-19" calendar="#gregorian">19 May</date>


But Humboldt and Ehrenberg (the latter being the author of the travel journal quoted) have been on the road some 180 days, so we either tag 2x180=360 dates, but still achieve no real information gain for the user, since @when would always be the same 180 times, or we invent a new value for @calendar combining two calendars that in this combination did not exist. Both alternatives seem unattractive to me.

And: The problem multiplies many times, since we have 10+ volumes of Humboldt's own journals still ahead of us, and since he is using and noting dates in at least three different calendars, the gregorian, the julian and the french revolutionary calendar. we would like to annotate which calendar(s) is/are used in the source, and give the 'normalised' gregorian date, but not necessarily want to tag every figure in notations like "7/19 May" with an extra <date>-tag.

In sum, we would prefer the efficient solution of havin only one <date>-element, but with multiple @calendar-values whereever applicable in the way Paul illustrated in his first example. Can we agree that @calendar should, as an option, be allowed to hold more than one calendar as its value?

Thank you all for your input, looking forward for more!
Best, Christian


 
pfs
 
 
 
On Wed, Aug 26, 2020, at 16:23, Martin Holmes wrote:
Hi Lou,
 
I think Christian was suggesting that @calendar permit multiple 
pointers, which I think makes perfect sense for scenarios like this. If 
you know of any good reasons not to make such a feature request, let us 
know; I think it would be a good idea myself.
 
Cheers,
Martin
 
On 2020-08-26 12:20 p.m., Lou Burnard wrote:
I am not sure what you are trying to achieve here. The calendar 
attribute permits only a single Uri. A Uri may begin with but not 
contain the # character. So whether or not you put spaces into it your 
suggested values are syntactically invalid.
 
On Wed, 26 Aug 2020, 19:13 Christian Thomas, 
<[hidden email] 
[hidden email]> wrote:
 
 
    Hi again, just to correct my own mistakes, I don't know where all
    the whitespaces went:
 
    > He was born on <date calendar="#gregorian #julian" when="1732-02-22">Feb. 22, 1732 (Feb. 11, 1731/32, O.S.</date>).
 
    Or, to take up Sascha Grabschs initial expample from the /edition
    humboldt digital/ (cf. https://edition-humboldt.de/H0016785/12r,
    currently tagged wrong, alas the question), "Dienstag 7/19 May"
    should be
 
    > <date when="1829-05-19" calendar="#gregorian #julian">Dienstag <hi rendition="#sup">7</hi>/<hi rendition="#sub">19</hi> May</date>
 
 
    All in favour?
 
    Best wishes
    Christian
 
 
    Am 26.08.2020 um 18:28 schrieb Martin Holmes:
    Unfortunately, @calendar currently allows only one pointer:
 
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
 
 
    but I think you should be able to use multiple pointers in this
    case; the use-case is very straightforward. Can anyone think of a
    reason why not?
 
    The same is not true of @datingMethod, though; because that
    specifies what system is being used for other attributes, and it's
    not possible to say which of the array of @*-custom attributes
    would use one dating method and which another, @datingMethod
    should stay as a single pointer.
 
    Cheers,
    Martin
 
    On 2020-08-26 9:20 a.m., Christian Thomas wrote:
    Dear TEI list, please excuse my maybe naive question, but, taking
    up the example from
    https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
    which says:
 
    He was born on <date calendar="#gregorian">Feb. 22,
    1732</date>(<date calendar="#julian" when="1732-02-22">Feb. 11,
    1731/32, O.S.</date>).
    Why can't I join the two (or more) attribute values in @calendar
    like this:
 
    He was born on <date calendar="#gregorian#julian"
    when="1732-02-22">Feb. 22, 1732(Feb. 11, 1731/32, O.S.</date>).
    This would (in my reading) say that in the source both calendars
    are being used, @when is the standardised Gregorian date as
    specified in the TEI Guidelines, and would save us the trouble of
    having to invent multiple new values for 'combined' calendars.
 
    Please excuse me if this is layed out as impossible resp.
    forbidden somewhere else (and, if so, thanks for pointing me to
    that place).
 
    Thanks, best wishes
    Christian
 
 
 
 
    Am 26.08.2020 um 17:19 schrieb Lou Burnard:
    The second part of your question is easy: the value of the
    @calendar attribute is a pointer to a definition of a calendar
    system. The definition might be provided by a <calendar> element
    inside a <calendarDesc> in your TEI header, or by some other
    location  entirely, like a Wikipedia entry.
 
    The first part is trickier. Does the @calendar reference apply
    to the value of @when, or to the content of the element? Hands
    up if your first thought (like mine) was the former! You're
    wrong. There's a warning note at
    https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
    which says clearly
    "the calendar attribute (unlike @datingMethod defined in
    att.datable.custom
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>)
    defines the calendar system of the date in the original material
    defined by the parent element, /not/ the calendar to which the
    date is normalized."
    So, if you want to indicate that the content of your date
    element is neither Julian nor Gregorian, but a combination of
    the two, you will have to define a new calendar (let's call it
    "julGreg") and reference that:
    <date calendar="#julGreg" when="1829-05-19">Dienstag 7/19
    May</date> , exactly as you propose. Note that @when is always
    understood to be normalised to the same (Gregorian) calendar. If
    you want to supply an additional normalisation to the Julian
    calendar, you'd have to add further attributes:
    <date calendar="#julGreg" when="1829-05-19"
    when-custom="1829-05-07" dating-method="#julian">Dienstag 7/19
    May</date>
 
    You will of course also need to provide somewhere a calendar
    element like this:
    <calendar xml:id="julGreg"><p>A composite calendar format in
    which both Julian and Gregorian forms are provided, separated by
    a solidus, with the Julian form first.</calendar>
 
 
 
    On Wed, 26 Aug 2020 at 12:09, Sascha Grabsch <[hidden email]
    [hidden email] [hidden email]
    [hidden email]> wrote:
 
        Dear list members,
 
        I have a question regarding the encoding of calendar systems
    using
        @calendar with <date>.
 
        Suppose we have a date given in a manuscript like this:
 
        > Dienstag 7/19 May
 
        This date is given in two calendar system, Julian and
    Gregorian.
        Would
        it be possible to tag it like this:
 
        <date calendar="#julian-gregorian"
    when="1829-05-19">Dienstag 7/19
        May</date>.
 
        My colleagues and I are unsure if this would be a valid
    usage of the
        <date> element, but I think this would be a clean way to
    mark up the
        given date (which in my reading is a _single_ date) and note
    the fact
        that it is given in two different calendar systems. If at all
        possible I
        would like to avoid having two <date> elements or *gasp*
    having to
        nest
        the <date> elements.
 
        We also are not sure what the URIs of @calendar should be
    pointing
        to in
        general. Are there defined concepts/authority file references
        available
        for calendar systems? I could only find examples of
    `#gregorian`,
        `#julian` in the TEI Guidelines.
 
        I would be very interested in your opinions and - if
    available -
        examples!
 
        All the best,
 
        Sascha Grabsch
 
    -- 
    Christian Thomas
    [hidden email]
    [hidden email]
    @dta_cthomas
    -- 
 
    -- 
    Christian Thomas
    [hidden email]  [hidden email]
    @dta_cthomas
    --
 
-- 
-------------------------------------
Humanities Computing and Media Centre
University of Victoria
[hidden email]
 

 

Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

Christian Thomas (HU Berlin)
In reply to this post by Kurz, Stephan

Dear Stephan, dear list,

Am 27.08.2020 um 10:33 schrieb Kurz, Stephan:

Dear Christian, dear list,

 

I could not think of any counter arguments apart from the solution not being immediately human-readable (only one normalized @when combined with two @calendar pointers).

Of course, one could explain this e.g. in calendarDesc; in the use case at hand with something like “Humboldt notes dates in both Julian and Gregorian variant, here @when only contains the normalized Gregorian date”. 


agreed, explicit info in the header always an advantage (for human readers, at least), but, my reading at least: @when by default and definition by TEI Guidelines always only refers to the normalized Gregorian date. So it would be clear anyway. This:

I guess adding the Julian date in @when-custom as I suggested yesterday would not provide additional information necessary and useful for a reader of today, even if it would add symmetry in having two calendars and two whens.


Plan is: We will not use @when-custom, since we will (to save resources and since e.g. the french revolutionary calendar (https://en.wikipedia.org/wiki/French_Republican_calendar) is not in use anymore and will not be queried by machines) not note the date as the value for @when-custom in a normalised notation "YYYY-MM-DD". We will also not use @datingMethod, since as stated above, @when 'automatically' sets the dating method to Gregorian, and so @datingMethod would only be necessary if you chose to date in your @when-custom attribute following a different calendar than the Gregorian. At least that would be my understanding and argumentation.

I can see that one might run into problems if you had @calendar="#julian #frenchRev" and wanted to normalise the given dates for both calendars using @when-custom and @datingMethod. But, for example (completely fictional, since at least we have not come across parallel usage of the Julian calendar, the dates of which Humboldt he noted parallel to the Gregorian ones while he was in Russia, and the french rev. calendar, which he noted when he was in France during the revolution):
<date when="1798-11-02" calendar="#julian #frenchRev" datingMethod="#julian" when-custom="1799-11-13">13 Feb jul./12 Brumaire</date>

would be fine and imo perfectly understandable (for humans & machines).

Does that make sense?

Best wishes
Christian


 

Best, Stephan

 

From: TEI (Text Encoding Initiative) public discussion list [hidden email] On Behalf Of Christian Thomas
Sent: Thursday, August 27, 2020 1:16 AM
To: [hidden email]
Subject: Re: Multiple values for @calendar in the same <date> element

 


Dear Martin, thank you:

On Wed, Aug 26, 2020, at 16:23, Martin Holmes wrote:
Hi Lou,
 
I think Christian was suggesting that @calendar permit multiple 
pointers, 

exactly what I meant.

Dear Paul,

Am 26.08.2020 um 22:51 schrieb Paul Schaffner:

Am I right in understanding this request to be aimed specifically
and exclusively at representations of date that more-or-less 
inextricably combine references to more than one calendar? Hence
 
 <date when="1829-05-19" calendar="#gregorian #julian">Thursday, 7/19 May</date>


yes, that's true and to be tagged like above, whereas this:


... whereas where the expression is separable, the need goes away?
 
 <date when="1829-05-19" calendar="#julian">Thursday, 7 May, O.S.</date> or
 <date when="1829-05-19" calendar="#gregorian">Thursday, 19 May, N.S.</date>


should be seperated. But this:


... and that the only alternative to this suggestion available within
the schema currently is to create a 'calendar' that is in effect a hybrid
between those referenced? like this?
 
 <date when="1829-05-19" calendar="#julian_gregorian">Thursday, 7/19 May</date>


is not the only anlternative, we could also go back to separating the date-Tags here, and we might have to, as long as date can only have one uri:

<date when="1829-05-19" calendar="#julian">Thursday, 7<date>/<date when="1829-05-19" calendar="#gregorian">19 May</date>


But Humboldt and Ehrenberg (the latter being the author of the travel journal quoted) have been on the road some 180 days, so we either tag 2x180=360 dates, but still achieve no real information gain for the user, since @when would always be the same 180 times, or we invent a new value for @calendar combining two calendars that in this combination did not exist. Both alternatives seem unattractive to me.

And: The problem multiplies many times, since we have 10+ volumes of Humboldt's own journals still ahead of us, and since he is using and noting dates in at least three different calendars, the gregorian, the julian and the french revolutionary calendar. we would like to annotate which calendar(s) is/are used in the source, and give the 'normalised' gregorian date, but not necessarily want to tag every figure in notations like "7/19 May" with an extra <date>-tag.

In sum, we would prefer the efficient solution of havin only one <date>-element, but with multiple @calendar-values whereever applicable in the way Paul illustrated in his first example. Can we agree that @calendar should, as an option, be allowed to hold more than one calendar as its value?

Thank you all for your input, looking forward for more!
Best, Christian


 
pfs
 
 
 
On Wed, Aug 26, 2020, at 16:23, Martin Holmes wrote:
Hi Lou,
 
I think Christian was suggesting that @calendar permit multiple 
pointers, which I think makes perfect sense for scenarios like this. If 
you know of any good reasons not to make such a feature request, let us 
know; I think it would be a good idea myself.
 
Cheers,
Martin
 
On 2020-08-26 12:20 p.m., Lou Burnard wrote:
I am not sure what you are trying to achieve here. The calendar 
attribute permits only a single Uri. A Uri may begin with but not 
contain the # character. So whether or not you put spaces into it your 
suggested values are syntactically invalid.
 
On Wed, 26 Aug 2020, 19:13 Christian Thomas, 
<[hidden email] 
[hidden email]> wrote:
 
 
    Hi again, just to correct my own mistakes, I don't know where all
    the whitespaces went:
 
    > He was born on <date calendar="#gregorian #julian" when="1732-02-22">Feb. 22, 1732 (Feb. 11, 1731/32, O.S.</date>).
 
    Or, to take up Sascha Grabschs initial expample from the /edition
    humboldt digital/ (cf. https://edition-humboldt.de/H0016785/12r,
    currently tagged wrong, alas the question), "Dienstag 7/19 May"
    should be
 
    > <date when="1829-05-19" calendar="#gregorian #julian">Dienstag <hi rendition="#sup">7</hi>/<hi rendition="#sub">19</hi> May</date>
 
 
    All in favour?
 
    Best wishes
    Christian
 
 
    Am 26.08.2020 um 18:28 schrieb Martin Holmes:
    Unfortunately, @calendar currently allows only one pointer:
 
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
 
 
    but I think you should be able to use multiple pointers in this
    case; the use-case is very straightforward. Can anyone think of a
    reason why not?
 
    The same is not true of @datingMethod, though; because that
    specifies what system is being used for other attributes, and it's
    not possible to say which of the array of @*-custom attributes
    would use one dating method and which another, @datingMethod
    should stay as a single pointer.
 
    Cheers,
    Martin
 
    On 2020-08-26 9:20 a.m., Christian Thomas wrote:
    Dear TEI list, please excuse my maybe naive question, but, taking
    up the example from
    https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
    which says:
 
    He was born on <date calendar="#gregorian">Feb. 22,
    1732</date>(<date calendar="#julian" when="1732-02-22">Feb. 11,
    1731/32, O.S.</date>).
    Why can't I join the two (or more) attribute values in @calendar
    like this:
 
    He was born on <date calendar="#gregorian#julian"
    when="1732-02-22">Feb. 22, 1732(Feb. 11, 1731/32, O.S.</date>).
    This would (in my reading) say that in the source both calendars
    are being used, @when is the standardised Gregorian date as
    specified in the TEI Guidelines, and would save us the trouble of
    having to invent multiple new values for 'combined' calendars.
 
    Please excuse me if this is layed out as impossible resp.
    forbidden somewhere else (and, if so, thanks for pointing me to
    that place).
 
    Thanks, best wishes
    Christian
 
 
 
 
    Am 26.08.2020 um 17:19 schrieb Lou Burnard:
    The second part of your question is easy: the value of the
    @calendar attribute is a pointer to a definition of a calendar
    system. The definition might be provided by a <calendar> element
    inside a <calendarDesc> in your TEI header, or by some other
    location  entirely, like a Wikipedia entry.
 
    The first part is trickier. Does the @calendar reference apply
    to the value of @when, or to the content of the element? Hands
    up if your first thought (like mine) was the former! You're
    wrong. There's a warning note at
    https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
    which says clearly
    "the calendar attribute (unlike @datingMethod defined in
    att.datable.custom
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>)
    defines the calendar system of the date in the original material
    defined by the parent element, /not/ the calendar to which the
    date is normalized."
    So, if you want to indicate that the content of your date
    element is neither Julian nor Gregorian, but a combination of
    the two, you will have to define a new calendar (let's call it
    "julGreg") and reference that:
    <date calendar="#julGreg" when="1829-05-19">Dienstag 7/19
    May</date> , exactly as you propose. Note that @when is always
    understood to be normalised to the same (Gregorian) calendar. If
    you want to supply an additional normalisation to the Julian
    calendar, you'd have to add further attributes:
    <date calendar="#julGreg" when="1829-05-19"
    when-custom="1829-05-07" dating-method="#julian">Dienstag 7/19
    May</date>
 
    You will of course also need to provide somewhere a calendar
    element like this:
    <calendar xml:id="julGreg"><p>A composite calendar format in
    which both Julian and Gregorian forms are provided, separated by
    a solidus, with the Julian form first.</calendar>
 
 
 
    On Wed, 26 Aug 2020 at 12:09, Sascha Grabsch <[hidden email]
    [hidden email] [hidden email]
    [hidden email]> wrote:
 
        Dear list members,
 
        I have a question regarding the encoding of calendar systems
    using
        @calendar with <date>.
 
        Suppose we have a date given in a manuscript like this:
 
        > Dienstag 7/19 May
 
        This date is given in two calendar system, Julian and
    Gregorian.
        Would
        it be possible to tag it like this:
 
        <date calendar="#julian-gregorian"
    when="1829-05-19">Dienstag 7/19
        May</date>.
 
        My colleagues and I are unsure if this would be a valid
    usage of the
        <date> element, but I think this would be a clean way to
    mark up the
        given date (which in my reading is a _single_ date) and note
    the fact
        that it is given in two different calendar systems. If at all
        possible I
        would like to avoid having two <date> elements or *gasp*
    having to
        nest
        the <date> elements.
 
        We also are not sure what the URIs of @calendar should be
    pointing
        to in
        general. Are there defined concepts/authority file references
        available
        for calendar systems? I could only find examples of
    `#gregorian`,
        `#julian` in the TEI Guidelines.
 
        I would be very interested in your opinions and - if
    available -
        examples!
 
        All the best,
 
        Sascha Grabsch
 
    -- 
    Christian Thomas
    [hidden email]
    [hidden email]
    @dta_cthomas
    -- 
 
    -- 
    Christian Thomas
    [hidden email]  [hidden email]
    @dta_cthomas
    --
 
-- 
-------------------------------------
Humanities Computing and Media Centre
University of Victoria
[hidden email]
 

 


-- 
Christian Thomas
[hidden email]
@dta_cthomas
--
Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

Christian Thomas (HU Berlin)
In reply to this post by Daniele Marotta-2

Dear Daniele,

Am 27.08.2020 um 10:43 schrieb Daniele Marotta:

Dear Stephan, dear list,

I'm afraid it would not be sufficient the @when-custom attribute if "he is using and noting dates in at least three different calendars, the gregorian, the julian and the french revolutionary calendar". I guess there could be cases where they have all three together.


it seems unlinkely, at least for our use case edition humboldt digital https://edition-humboldt.de/, that all three dates are noted in the source, but if they were, I would still survive without @when-custom:
<date when="1798-11-02" calendar="#gregorian #julian #frenchRev">2 Feb greg./13 Feb jul./12 Brumaire</date>


Or did I miss your point? If so, please excuse me and explain a little further. Thanks!

Best wishes, Christian



Best, Daniele

Il 27/08/2020 10:33, Kurz, Stephan ha scritto:

Dear Christian, dear list,

 

I could not think of any counter arguments apart from the solution not being immediately human-readable (only one normalized @when combined with two @calendar pointers).

Of course, one could explain this e.g. in calendarDesc; in the use case at hand with something like “Humboldt notes dates in both Julian and Gregorian variant, here @when only contains the normalized Gregorian date”.  

I guess adding the Julian date in @when-custom as I suggested yesterday would not provide additional information necessary and useful for a reader of today, even if it would add symmetry in having two calendars and two whens.

 

Best, Stephan

 

From: TEI (Text Encoding Initiative) public discussion list [hidden email] On Behalf Of Christian Thomas
Sent: Thursday, August 27, 2020 1:16 AM
To: [hidden email]
Subject: Re: Multiple values for @calendar in the same <date> element

 


Dear Martin, thank you:

On Wed, Aug 26, 2020, at 16:23, Martin Holmes wrote:
Hi Lou,
 
I think Christian was suggesting that @calendar permit multiple 
pointers, 

exactly what I meant.

Dear Paul,

Am 26.08.2020 um 22:51 schrieb Paul Schaffner:

Am I right in understanding this request to be aimed specifically
and exclusively at representations of date that more-or-less 
inextricably combine references to more than one calendar? Hence
 
 <date when="1829-05-19" calendar="#gregorian #julian">Thursday, 7/19 May</date>


yes, that's true and to be tagged like above, whereas this:


... whereas where the expression is separable, the need goes away?
 
 <date when="1829-05-19" calendar="#julian">Thursday, 7 May, O.S.</date> or
 <date when="1829-05-19" calendar="#gregorian">Thursday, 19 May, N.S.</date>


should be seperated. But this:


... and that the only alternative to this suggestion available within
the schema currently is to create a 'calendar' that is in effect a hybrid
between those referenced? like this?
 
 <date when="1829-05-19" calendar="#julian_gregorian">Thursday, 7/19 May</date>


is not the only anlternative, we could also go back to separating the date-Tags here, and we might have to, as long as date can only have one uri:

<date when="1829-05-19" calendar="#julian">Thursday, 7<date>/<date when="1829-05-19" calendar="#gregorian">19 May</date>


But Humboldt and Ehrenberg (the latter being the author of the travel journal quoted) have been on the road some 180 days, so we either tag 2x180=360 dates, but still achieve no real information gain for the user, since @when would always be the same 180 times, or we invent a new value for @calendar combining two calendars that in this combination did not exist. Both alternatives seem unattractive to me.

And: The problem multiplies many times, since we have 10+ volumes of Humboldt's own journals still ahead of us, and since he is using and noting dates in at least three different calendars, the gregorian, the julian and the french revolutionary calendar. we would like to annotate which calendar(s) is/are used in the source, and give the 'normalised' gregorian date, but not necessarily want to tag every figure in notations like "7/19 May" with an extra <date>-tag.

In sum, we would prefer the efficient solution of havin only one <date>-element, but with multiple @calendar-values whereever applicable in the way Paul illustrated in his first example. Can we agree that @calendar should, as an option, be allowed to hold more than one calendar as its value?

Thank you all for your input, looking forward for more!
Best, Christian


 
pfs
 
 
 
On Wed, Aug 26, 2020, at 16:23, Martin Holmes wrote:
Hi Lou,
 
I think Christian was suggesting that @calendar permit multiple 
pointers, which I think makes perfect sense for scenarios like this. If 
you know of any good reasons not to make such a feature request, let us 
know; I think it would be a good idea myself.
 
Cheers,
Martin
 
On 2020-08-26 12:20 p.m., Lou Burnard wrote:
I am not sure what you are trying to achieve here. The calendar 
attribute permits only a single Uri. A Uri may begin with but not 
contain the # character. So whether or not you put spaces into it your 
suggested values are syntactically invalid.
 
On Wed, 26 Aug 2020, 19:13 Christian Thomas, 
<[hidden email] 
[hidden email]> wrote:
 
 
    Hi again, just to correct my own mistakes, I don't know where all
    the whitespaces went:
 
    > He was born on <date calendar="#gregorian #julian" when="1732-02-22">Feb. 22, 1732 (Feb. 11, 1731/32, O.S.</date>).
 
    Or, to take up Sascha Grabschs initial expample from the /edition
    humboldt digital/ (cf. https://edition-humboldt.de/H0016785/12r,
    currently tagged wrong, alas the question), "Dienstag 7/19 May"
    should be
 
    > <date when="1829-05-19" calendar="#gregorian #julian">Dienstag <hi rendition="#sup">7</hi>/<hi rendition="#sub">19</hi> May</date>
 
 
    All in favour?
 
    Best wishes
    Christian
 
 
    Am 26.08.2020 um 18:28 schrieb Martin Holmes:
    Unfortunately, @calendar currently allows only one pointer:
 
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
 
 
    but I think you should be able to use multiple pointers in this
    case; the use-case is very straightforward. Can anyone think of a
    reason why not?
 
    The same is not true of @datingMethod, though; because that
    specifies what system is being used for other attributes, and it's
    not possible to say which of the array of @*-custom attributes
    would use one dating method and which another, @datingMethod
    should stay as a single pointer.
 
    Cheers,
    Martin
 
    On 2020-08-26 9:20 a.m., Christian Thomas wrote:
    Dear TEI list, please excuse my maybe naive question, but, taking
    up the example from
    https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
    which says:
 
    He was born on <date calendar="#gregorian">Feb. 22,
    1732</date>(<date calendar="#julian" when="1732-02-22">Feb. 11,
    1731/32, O.S.</date>).
    Why can't I join the two (or more) attribute values in @calendar
    like this:
 
    He was born on <date calendar="#gregorian#julian"
    when="1732-02-22">Feb. 22, 1732(Feb. 11, 1731/32, O.S.</date>).
    This would (in my reading) say that in the source both calendars
    are being used, @when is the standardised Gregorian date as
    specified in the TEI Guidelines, and would save us the trouble of
    having to invent multiple new values for 'combined' calendars.
 
    Please excuse me if this is layed out as impossible resp.
    forbidden somewhere else (and, if so, thanks for pointing me to
    that place).
 
    Thanks, best wishes
    Christian
 
 
 
 
    Am 26.08.2020 um 17:19 schrieb Lou Burnard:
    The second part of your question is easy: the value of the
    @calendar attribute is a pointer to a definition of a calendar
    system. The definition might be provided by a <calendar> element
    inside a <calendarDesc> in your TEI header, or by some other
    location  entirely, like a Wikipedia entry.
 
    The first part is trickier. Does the @calendar reference apply
    to the value of @when, or to the content of the element? Hands
    up if your first thought (like mine) was the former! You're
    wrong. There's a warning note at
    https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
    which says clearly
    "the calendar attribute (unlike @datingMethod defined in
    att.datable.custom
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>)
    defines the calendar system of the date in the original material
    defined by the parent element, /not/ the calendar to which the
    date is normalized."
    So, if you want to indicate that the content of your date
    element is neither Julian nor Gregorian, but a combination of
    the two, you will have to define a new calendar (let's call it
    "julGreg") and reference that:
    <date calendar="#julGreg" when="1829-05-19">Dienstag 7/19
    May</date> , exactly as you propose. Note that @when is always
    understood to be normalised to the same (Gregorian) calendar. If
    you want to supply an additional normalisation to the Julian
    calendar, you'd have to add further attributes:
    <date calendar="#julGreg" when="1829-05-19"
    when-custom="1829-05-07" dating-method="#julian">Dienstag 7/19
    May</date>
 
    You will of course also need to provide somewhere a calendar
    element like this:
    <calendar xml:id="julGreg"><p>A composite calendar format in
    which both Julian and Gregorian forms are provided, separated by
    a solidus, with the Julian form first.</calendar>
 
 
 
    On Wed, 26 Aug 2020 at 12:09, Sascha Grabsch <[hidden email]
    [hidden email] [hidden email]
    [hidden email]> wrote:
 
        Dear list members,
 
        I have a question regarding the encoding of calendar systems
    using
        @calendar with <date>.
 
        Suppose we have a date given in a manuscript like this:
 
        > Dienstag 7/19 May
 
        This date is given in two calendar system, Julian and
    Gregorian.
        Would
        it be possible to tag it like this:
 
        <date calendar="#julian-gregorian"
    when="1829-05-19">Dienstag 7/19
        May</date>.
 
        My colleagues and I are unsure if this would be a valid
    usage of the
        <date> element, but I think this would be a clean way to
    mark up the
        given date (which in my reading is a _single_ date) and note
    the fact
        that it is given in two different calendar systems. If at all
        possible I
        would like to avoid having two <date> elements or *gasp*
    having to
        nest
        the <date> elements.
 
        We also are not sure what the URIs of @calendar should be
    pointing
        to in
        general. Are there defined concepts/authority file references
        available
        for calendar systems? I could only find examples of
    `#gregorian`,
        `#julian` in the TEI Guidelines.
 
        I would be very interested in your opinions and - if
    available -
        examples!
 
        All the best,
 
        Sascha Grabsch
 
    -- 
    Christian Thomas
    [hidden email]
    [hidden email]
    @dta_cthomas
    -- 
 
    -- 
    Christian Thomas
    [hidden email]  [hidden email]
    @dta_cthomas
    --
 
-- 
-------------------------------------
Humanities Computing and Media Centre
University of Victoria
[hidden email]
 

 


-- 
Christian Thomas
[hidden email]
@dta_cthomas
--
lou
Reply | Threaded
Open this post in threaded view
|

Re: Multiple values for @calendar in the same <date> element

lou
In reply to this post by Thibault Clérice
I think Martin suggested this approach at the beginning of the thread. He characterized it as "gnarly" ... a curious word but not an entirely positive one.

On Thu, 27 Aug 2020, 11:10 Thibault Clérice, <[hidden email]> wrote:

Dear all,

This is probably my first message here, but I have not seen the following proposal:

He was born on <date calendar="#gregorian">Feb. 22, 1732(<date calendar="#julian">Feb. 11, 1731/32, O.S.</date>).</date>
 <date when="1829-05-19" calendar="#gregorian>Thursday, 7<date calendar= #julian">/19</date> May</date>

This, or another tag to be fair, would allow to no duplicate the @when while keeping the information about the different calendar. The issue with both pointer is that you do not define which segment is a representation of which calendar system. Here, with a tree-like solution (rather than sequential nodes that would meet many exceptions), we could actually clear things up ?

If this had been brought up, I am sorry, I did not find it in the thread...

Best,

Thibault


On 27/08/2020 10:43, Daniele Marotta wrote:

Dear Stephan, dear list,

I'm afraid it would not be sufficient the @when-custom attribute if "he is using and noting dates in at least three different calendars, the gregorian, the julian and the french revolutionary calendar". I guess there could be cases where they have all three together.

Best, Daniele

Il 27/08/2020 10:33, Kurz, Stephan ha scritto:

Dear Christian, dear list,

 

I could not think of any counter arguments apart from the solution not being immediately human-readable (only one normalized @when combined with two @calendar pointers).

Of course, one could explain this e.g. in calendarDesc; in the use case at hand with something like “Humboldt notes dates in both Julian and Gregorian variant, here @when only contains the normalized Gregorian date”.  

I guess adding the Julian date in @when-custom as I suggested yesterday would not provide additional information necessary and useful for a reader of today, even if it would add symmetry in having two calendars and two whens.

 

Best, Stephan

 

From: TEI (Text Encoding Initiative) public discussion list [hidden email] On Behalf Of Christian Thomas
Sent: Thursday, August 27, 2020 1:16 AM
To: [hidden email]
Subject: Re: Multiple values for @calendar in the same <date> element

 


Dear Martin, thank you:

On Wed, Aug 26, 2020, at 16:23, Martin Holmes wrote:
Hi Lou,
 
I think Christian was suggesting that @calendar permit multiple 
pointers, 

exactly what I meant.

Dear Paul,

Am 26.08.2020 um 22:51 schrieb Paul Schaffner:

Am I right in understanding this request to be aimed specifically
and exclusively at representations of date that more-or-less 
inextricably combine references to more than one calendar? Hence
 
 <date when="1829-05-19" calendar="#gregorian #julian">Thursday, 7/19 May</date>


yes, that's true and to be tagged like above, whereas this:


... whereas where the expression is separable, the need goes away?
 
 <date when="1829-05-19" calendar="#julian">Thursday, 7 May, O.S.</date> or
 <date when="1829-05-19" calendar="#gregorian">Thursday, 19 May, N.S.</date>


should be seperated. But this:


... and that the only alternative to this suggestion available within
the schema currently is to create a 'calendar' that is in effect a hybrid
between those referenced? like this?
 
 <date when="1829-05-19" calendar="#julian_gregorian">Thursday, 7/19 May</date>


is not the only anlternative, we could also go back to separating the date-Tags here, and we might have to, as long as date can only have one uri:

<date when="1829-05-19" calendar="#julian">Thursday, 7<date>/<date when="1829-05-19" calendar="#gregorian">19 May</date>


But Humboldt and Ehrenberg (the latter being the author of the travel journal quoted) have been on the road some 180 days, so we either tag 2x180=360 dates, but still achieve no real information gain for the user, since @when would always be the same 180 times, or we invent a new value for @calendar combining two calendars that in this combination did not exist. Both alternatives seem unattractive to me.

And: The problem multiplies many times, since we have 10+ volumes of Humboldt's own journals still ahead of us, and since he is using and noting dates in at least three different calendars, the gregorian, the julian and the french revolutionary calendar. we would like to annotate which calendar(s) is/are used in the source, and give the 'normalised' gregorian date, but not necessarily want to tag every figure in notations like "7/19 May" with an extra <date>-tag.

In sum, we would prefer the efficient solution of havin only one <date>-element, but with multiple @calendar-values whereever applicable in the way Paul illustrated in his first example. Can we agree that @calendar should, as an option, be allowed to hold more than one calendar as its value?

Thank you all for your input, looking forward for more!
Best, Christian


 
pfs
 
 
 
On Wed, Aug 26, 2020, at 16:23, Martin Holmes wrote:
Hi Lou,
 
I think Christian was suggesting that @calendar permit multiple 
pointers, which I think makes perfect sense for scenarios like this. If 
you know of any good reasons not to make such a feature request, let us 
know; I think it would be a good idea myself.
 
Cheers,
Martin
 
On 2020-08-26 12:20 p.m., Lou Burnard wrote:
I am not sure what you are trying to achieve here. The calendar 
attribute permits only a single Uri. A Uri may begin with but not 
contain the # character. So whether or not you put spaces into it your 
suggested values are syntactically invalid.
 
On Wed, 26 Aug 2020, 19:13 Christian Thomas, 
<[hidden email] 
[hidden email]> wrote:
 
 
    Hi again, just to correct my own mistakes, I don't know where all
    the whitespaces went:
 
    > He was born on <date calendar="#gregorian #julian" when="1732-02-22">Feb. 22, 1732 (Feb. 11, 1731/32, O.S.</date>).
 
    Or, to take up Sascha Grabschs initial expample from the /edition
    humboldt digital/ (cf. https://edition-humboldt.de/H0016785/12r,
    currently tagged wrong, alas the question), "Dienstag 7/19 May"
    should be
 
    > <date when="1829-05-19" calendar="#gregorian #julian">Dienstag <hi rendition="#sup">7</hi>/<hi rendition="#sub">19</hi> May</date>
 
 
    All in favour?
 
    Best wishes
    Christian
 
 
    Am 26.08.2020 um 18:28 schrieb Martin Holmes:
    Unfortunately, @calendar currently allows only one pointer:
 
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html>
 
 
    but I think you should be able to use multiple pointers in this
    case; the use-case is very straightforward. Can anyone think of a
    reason why not?
 
    The same is not true of @datingMethod, though; because that
    specifies what system is being used for other attributes, and it's
    not possible to say which of the array of @*-custom attributes
    would use one dating method and which another, @datingMethod
    should stay as a single pointer.
 
    Cheers,
    Martin
 
    On 2020-08-26 9:20 a.m., Christian Thomas wrote:
    Dear TEI list, please excuse my maybe naive question, but, taking
    up the example from
    https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
    which says:
 
    He was born on <date calendar="#gregorian">Feb. 22,
    1732</date>(<date calendar="#julian" when="1732-02-22">Feb. 11,
    1731/32, O.S.</date>).
    Why can't I join the two (or more) attribute values in @calendar
    like this:
 
    He was born on <date calendar="#gregorian#julian"
    when="1732-02-22">Feb. 22, 1732(Feb. 11, 1731/32, O.S.</date>).
    This would (in my reading) say that in the source both calendars
    are being used, @when is the standardised Gregorian date as
    specified in the TEI Guidelines, and would save us the trouble of
    having to invent multiple new values for 'combined' calendars.
 
    Please excuse me if this is layed out as impossible resp.
    forbidden somewhere else (and, if so, thanks for pointing me to
    that place).
 
    Thanks, best wishes
    Christian
 
 
 
 
    Am 26.08.2020 um 17:19 schrieb Lou Burnard:
    The second part of your question is easy: the value of the
    @calendar attribute is a pointer to a definition of a calendar
    system. The definition might be provided by a <calendar> element
    inside a <calendarDesc> in your TEI header, or by some other
    location  entirely, like a Wikipedia entry.
 
    The first part is trickier. Does the @calendar reference apply
    to the value of @when, or to the content of the element? Hands
    up if your first thought (like mine) was the former! You're
    wrong. There's a warning note at
    https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.html
    which says clearly
    "the calendar attribute (unlike @datingMethod defined in
    att.datable.custom
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>
    <https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-att.datable.custom.html>)
    defines the calendar system of the date in the original material
    defined by the parent element, /not/ the calendar to which the
    date is normalized."
    So, if you want to indicate that the content of your date
    element is neither Julian nor Gregorian, but a combination of
    the two, you will have to define a new calendar (let's call it
    "julGreg") and reference that:
    <date calendar="#julGreg" when="1829-05-19">Dienstag 7/19
    May</date> , exactly as you propose. Note that @when is always
    understood to be normalised to the same (Gregorian) calendar. If
    you want to supply an additional normalisation to the Julian
    calendar, you'd have to add further attributes:
    <date calendar="#julGreg" when="1829-05-19"
    when-custom="1829-05-07" dating-method="#julian">Dienstag 7/19
    May</date>
 
    You will of course also need to provide somewhere a calendar
    element like this:
    <calendar xml:id="julGreg"><p>A composite calendar format in
    which both Julian and Gregorian forms are provided, separated by
    a solidus, with the Julian form first.</calendar>
 
 
 
    On Wed, 26 Aug 2020 at 12:09, Sascha Grabsch <[hidden email]
    [hidden email] [hidden email]
    [hidden email]> wrote:
 
        Dear list members,
 
        I have a question regarding the encoding of calendar systems
    using
        @calendar with <date>.
 
        Suppose we have a date given in a manuscript like this:
 
        > Dienstag 7/19 May
 
        This date is given in two calendar system, Julian and
    Gregorian.
        Would
        it be possible to tag it like this:
 
        <date calendar="#julian-gregorian"
    when="1829-05-19">Dienstag 7/19
        May</date>.
 
        My colleagues and I are unsure if this would be a valid
    usage of the
        <date> element, but I think this would be a clean way to
    mark up the
        given date (which in my reading is a _single_ date) and note
    the fact
        that it is given in two different calendar systems. If at all
        possible I
        would like to avoid having two <date> elements or *gasp*
    having to
        nest
        the <date> elements.
 
        We also are not sure what the URIs of @calendar should be
    pointing
        to in
        general. Are there defined concepts/authority file references
        available
        for calendar systems? I could only find examples of
    `#gregorian`,
        `#julian` in the TEI Guidelines.
 
        I would be very interested in your opinions and - if
    available -
        examples!
 
        All the best,
 
        Sascha Grabsch
 
    -- 
    Christian Thomas
    [hidden email]
    [hidden email]
    @dta_cthomas
    -- 
 
    -- 
    Christian Thomas
    [hidden email]  [hidden email]
    @dta_cthomas
    --
 
-- 
-------------------------------------
Humanities Computing and Media Centre
University of Victoria
[hidden email]
 

 

12