Formal Public Identifiers in P5 DTDs

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

Formal Public Identifiers in P5 DTDs

Martin Holmes
Hi all,

The P5 build system currently makes use of Formal Public Identifiers for
TEI modules which are encoded in <altIdent type="FPI"> in the TEI
<moduleSpec> elements; the only place these are used is in the generated
TEI DTDs such as the base TEI DTD:

<https://tei-c.org/Vault/P5/current/xml/tei/schema/dtd/tei.dtd>

where you will see things such as:

<![%TEI.transcr;[
        <!ENTITY % file.transcr PUBLIC '-//TEI P5//ELEMENTS Transcription of
Primary Sources//EN' 'transcr.dtd' >
        %file.transcr;
        ]]>

Is there anyone out there who is making use of these FPIs, or are they
essentially redundant?

Cheers,
Martin

--
------------------------------------------
Martin Holmes
UVic Humanities Computing and Media Centre
Reply | Threaded
Open this post in threaded view
|

Re: Formal Public Identifiers in P5 DTDs

C. M. Sperberg-McQueen
> On 13,Nov2020, at 10:50 AM, Martin Holmes <[hidden email]> wrote:
>
> Hi all,
>
> The P5 build system currently makes use of Formal Public Identifiers for TEI modules which are encoded in <altIdent type="FPI"> in the TEI <moduleSpec> elements; the only place these are used is in the generated TEI DTDs such as the base TEI DTD:
>
> <https://tei-c.org/Vault/P5/current/xml/tei/schema/dtd/tei.dtd>
>
> where you will see things such as:
>
> <![%TEI.transcr;[
> <!ENTITY % file.transcr PUBLIC '-//TEI P5//ELEMENTS Transcription of Primary Sources//EN' 'transcr.dtd' >
> %file.transcr;
> ]]>
>
> Is there anyone out there who is making use of these FPIs, or are they essentially redundant?

I am not currently making use of the FPIs you refer to.  When I write XML catalogs, I use formal public identifiers as the match patterns wherever possible, so losing the public identifiers, or making them no longer formal public identifiers, would feel like a definite step backwards to me.  But I may be an outlier: there are some parts of the XML infrastructure that I use a lot that others appear to make no use of at all; FPIs may be another.

Like other parts of the basic parsing and processing infrastructure, catalogs may be worked with intensively for a short time until the setup is working and then be forgotten until it is necessary to adjust them again.  That leads me to suspect that there may be people whose setups make use of these FPIs without their being consciously aware of the fact.  

My two cents.  Take only as directed; if nausea persists consult a qualified technologist or voudou practitioner.

Michael

********************************************
C. M. Sperberg-McQueen
Black Mesa Technologies LLC
[hidden email]
http://www.blackmesatech.com
********************************************
Reply | Threaded
Open this post in threaded view
|

Re: Formal Public Identifiers in P5 DTDs

Peter Flynn-8
On 13/11/2020 20:35, C. M. Sperberg-McQueen wrote:

>> On 13,Nov2020, at 10:50 AM, Martin Holmes <[hidden email]> wrote:
>>
>> The P5 build system currently makes use of Formal Public
>> Identifiers [...] the only place these are used is in the generated
>> TEI DTDs such as the base TEI DTD:
>> [...]
>> Is there anyone out there who is making use of these FPIs, or are
>> they essentially redundant?
>
> I am not currently making use of the FPIs you refer to.  When I write
> XML catalogs, I use formal public identifiers as the match patterns
> wherever possible, so losing the public identifiers, or making them
> no longer formal public identifiers, would feel like a definite step
> backwards to me.  But I may be an outlier: there are some parts of
> the XML infrastructure that I use a lot that others appear to make no
> use of at all; FPIs may be another.

I also use them in exactly the same way, so no, you're not an outlier;
or we both are. However, knowing that they may disappear means I can at
least plan for the future.

> Like other parts of the basic parsing and processing infrastructure,
> catalogs may be worked with intensively for a short time until the
> setup is working and then be forgotten until it is necessary to
> adjust them again.  That leads me to suspect that there may be
> people whose setups make use of these FPIs without their being
> consciously aware of the fact.

That is entirely possible, especially where a functioning project
undertakes XML TEI encoding in an environment established long ago. and
not modified since, because it "just works", and the people who set it
up are long since moved on.

I am a fan of FPIs in one situation only: the document type declaration
at the top of an XML document. As a company which is a registered public
owner within the meaning of ISO/IEC 9070:1991, we lay claim to the
ownership of the identifier for entities over which we have change-rights.

> My two cents.  Take only as directed; if nausea persists consult a
> qualified technologist or voudou practitioner.

I did but she said I was on drugs.

P
lou
Reply | Threaded
Open this post in threaded view
|

Re: Formal Public Identifiers in P5 DTDs

lou
In reply to this post by Martin Holmes
Is this the thin end of a wedge designed to lead to withdrawal of support for DTDs in general?  I can't see any point in picking on one particular part of the DTD support system, unless you want to cripple it gradually so that puting the poor old horse down looks like an act of mercy.



On Fri, 13 Nov 2020 at 17:50, Martin Holmes <[hidden email]> wrote:
Hi all,

The P5 build system currently makes use of Formal Public Identifiers for
TEI modules which are encoded in <altIdent type="FPI"> in the TEI
<moduleSpec> elements; the only place these are used is in the generated
TEI DTDs such as the base TEI DTD:

<https://tei-c.org/Vault/P5/current/xml/tei/schema/dtd/tei.dtd>

where you will see things such as:

<![%TEI.transcr;[
        <!ENTITY % file.transcr PUBLIC '-//TEI P5//ELEMENTS Transcription of
Primary Sources//EN' 'transcr.dtd' >
        %file.transcr;
        ]]>

Is there anyone out there who is making use of these FPIs, or are they
essentially redundant?

Cheers,
Martin

--
------------------------------------------
Martin Holmes
UVic Humanities Computing and Media Centre
Reply | Threaded
Open this post in threaded view
|

Re: Formal Public Identifiers in P5 DTDs

Martin Holmes
Hi Lou,

The reason for the query was this ticket:

<https://github.com/TEIC/TEI/issues/2049>

which is about the horrible way we currently encode the strings which
become FPIs, using <altIdent>. In talking about what might make a better
approach to encoding them, we're of course obliged to consider the
possibility that not having them at all might make life simpler.

As you know, I would dearly love to see the back of DTD support, but
that's not actually what this is about.

Cheers,
Martin

On 2020-11-13 2:42 p.m., Lou Burnard wrote:

> Is this the thin end of a wedge designed to lead to withdrawal of
> support for DTDs in general?  I can't see any point in picking on one
> particular part of the DTD support system, unless you want to cripple it
> gradually so that puting the poor old horse down looks like an act of
> mercy.
>
>
>
> On Fri, 13 Nov 2020 at 17:50, Martin Holmes <[hidden email]
> <mailto:[hidden email]>> wrote:
>
>     Hi all,
>
>     The P5 build system currently makes use of Formal Public Identifiers
>     for
>     TEI modules which are encoded in <altIdent type="FPI"> in the TEI
>     <moduleSpec> elements; the only place these are used is in the
>     generated
>     TEI DTDs such as the base TEI DTD:
>
>     <https://tei-c.org/Vault/P5/current/xml/tei/schema/dtd/tei.dtd>
>
>     where you will see things such as:
>
>     <![%TEI.transcr;[
>              <!ENTITY % file.transcr PUBLIC '-//TEI P5//ELEMENTS
>     Transcription of
>     Primary Sources//EN' 'transcr.dtd' >
>              %file.transcr;
>              ]]>
>
>     Is there anyone out there who is making use of these FPIs, or are they
>     essentially redundant?
>
>     Cheers,
>     Martin
>
>     --
>     ------------------------------------------
>     Martin Holmes
>     UVic Humanities Computing and Media Centre
>

--
------------------------------------------
Martin Holmes
UVic Humanities Computing and Media Centre
Reply | Threaded
Open this post in threaded view
|

Re: Formal Public Identifiers in P5 DTDs

Bauman, Syd
In reply to this post by lou
Ha! Thank you, Lou; I laughed out loud.

This idea was not designed to lead to the demise of DTDs in TEI, but when Martin and I spoke about it, we cheerfully conceded the point that it may help speed up that demise.

—Syd, old horse crippler.

P.S. I lean towards being very cautious about removing the FPIs for the reasons Michael and Peter have mentioned: folks that don’t really know they are using them in a catalog file. I figure most people can look to see if they are used in their DOCTYPE declaration or not, but system catalog files may be more mysterious. That said, I spent over ¼ hour today looking for such catalog files on my own system — I know I used to use them — but could not find any that mentioned the TEI P5 FPIs.


Is this the thin end of a wedge designed to lead to withdrawal of support for DTDs in general?  I can't see any point in picking on one particular part of the DTD support system, unless you want to cripple it gradually so that puting the poor old horse down looks like an act of mercy.