workflow support in TEI?

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

workflow support in TEI?

Torsten Roeder
Dear TEI-L,

I was wondering to what extent (if at all) TEI supports project
workflows. To give an example: Let us assume I want to keep track of all
<bibl> elements in <teiHeader> to see what their current status is,
regarding the quality of information and the research done so far.
Sometimes there is only raw data that was imported or copy/pasted by
editors from catalogues which still need to be proofread.

What I have tried so far to approach this task:

* Working with XML comments. This seems manageable when only one or very
few editors (=two) are involved. Otherwise it has a good chance to
result in chaos.

* Misusing @cert. This works not that badly, but I am still not happy
about it, because @cert refers to the quality of an intervention or an
interpretation, and not to the status of editing workflow I intended.

* Editing my .rnc file, define a project namespace, then define a new
@status attribute (something like @my:status), and allow some project
specific workflow terms. For now, this is my preferred option.

My questions to you, dear TEI-L: Do you see any other options within or
outside the TEI? In general, I see this as kind of a quality management
layer, which sould not be necessarily a part of TEI, but would be of
great interest for other projects. I am sure that I am not the first
person that has spent some thoughts on this issue.

Best regards,
Torsten


--
Torsten Roeder
Wissenschaftlicher Mitarbeiter
Richard Wagner Schriften

Julius-Maximilians-Universität Würzburg
Institut für Musikforschung
Domerschulstraße 13
D-97070 Würzburg

Tel  +49 (0)931 31-85167
Mail [hidden email]
WWW  http://www.musikwissenschaft.uni-wuerzburg.de/rws
Reply | Threaded
Open this post in threaded view
|

Re: workflow support in TEI?

James Cummings-4
On 17/08/17 09:56, Torsten Roeder wrote:
> Dear TEI-L,
>

Hi Torsten,

> I was wondering to what extent (if at all) TEI supports project
> workflows. To give an example: Let us assume I want to keep
> track of all <bibl> elements in <teiHeader> to see what their
> current status is, regarding the quality of information and the
> research done so far. Sometimes there is only raw data that was
> imported or copy/pasted by editors from catalogues which still
> need to be proofread.

And I assume you want to do this on a per-<bibl> basis. i.e. not
just store than they've got up  to bibl123 but that bibl5 has
been updated, bibl6 is pending review, bibl7 has been
cut-and-pasted only, but bibl8 is entirely done.

> What I have tried so far to approach this task:
>
> * Working with XML comments. This seems manageable when only
> one or very few editors (=two) are involved. Otherwise it has a
> good chance to result in chaos.

While that is fine for in-project notes, it isn't a very rigorous
system.

> * Misusing @cert. This works not that badly, but I am still not
> happy about it, because @cert refers to the quality of an
> intervention or an interpretation, and not to the status of
> editing workflow I intended.

@cert definitely seems wrong.

> * Editing my .rnc file, define a project namespace, then define
> a new @status attribute (something like @my:status), and allow
> some project specific workflow terms. For now, this is my
> preferred option.

You do *not* do that by editing your .rnc file. Please, please do
not do that. Adding a new attribute, in a new namespace is fine,
but the TEI has a whole customization mechanism designed to help
you constrain and extend the TEI as needed. Your TEI ODD
customization file should do that (or via interfaces like Roma).
Do not hand edit the output schemas, they are output from the
customization file. All uses of TEI are a use of a customized
version of TEI (even tei_all).  We don't know if you are using a
particular popular customization, or one you've already made.


> My questions to you, dear TEI-L: Do you see any other options
> within or outside the TEI? In general, I see this as kind of a
> quality management layer, which sould not be necessarily a part
> of TEI, but would be of great interest for other projects. I am
> sure that I am not the first person that has spent some
> thoughts on this issue.

If this was general status of the file then @status on
<revisionDesc> is sometimes used.  (i.e. what milestone is the
file as a whole at).  Otherwise sometimes people use individual
changes inside <revisionDesc> to signal some stages in revision,
but that is really meant to record the revisions which have been
made. However, this is often used for project workflow purposes.

Since you want to track individual <bibl> why not use @status
already on this?  As you can see from
http://www.tei-c.org/release/doc/tei-p5-doc/en/html/ref-bibl.html 
the <bibl> element has the @status attribute available and if
your <bibl> elements are going through a particular set of stages
you can customize the TEI (using an ODD or via Roma) to only
allow certain values on this. 'proofed', 'unchecked', 'updated'
or what have you.

Hope that helps,
James



--

Dr James Cummings, [hidden email]
Academic IT Services, University of Oxford
Reply | Threaded
Open this post in threaded view
|

Re: workflow support in TEI?

Torsten Roeder
Hi James!

>> I was wondering to what extent (if at all) TEI supports project
>> workflows. To give an example: Let us assume I want to keep track of
>> all <bibl> elements in <teiHeader> to see what their current status
>> is, regarding the quality of information and the research done so far.
>> Sometimes there is only raw data that was imported or copy/pasted by
>> editors from catalogues which still need to be proofread.
>
> And I assume you want to do this on a per-<bibl> basis. i.e. not just
> store than they've got up  to bibl123 but that bibl5 has been updated,
> bibl6 is pending review, bibl7 has been cut-and-pasted only, but bibl8
> is entirely done.

Exactly, per-element basis. <bibl> is just one of more examples I could
give.

>> What I have tried so far to approach this task:
>>
>> * Working with XML comments. This seems manageable when only one or
>> very few editors (=two) are involved. Otherwise it has a good chance
>> to result in chaos.
>
> While that is fine for in-project notes, it isn't a very rigorous system.
>
>> * Misusing @cert. This works not that badly, but I am still not happy
>> about it, because @cert refers to the quality of an intervention or an
>> interpretation, and not to the status of editing workflow I intended.
>
> @cert definitely seems wrong.
>
>> * Editing my .rnc file, define a project namespace, then define a new
>> @status attribute (something like @my:status), and allow some project
>> specific workflow terms. For now, this is my preferred option.
>
> You do *not* do that by editing your .rnc file. Please, please do not do
> that. Adding a new attribute, in a new namespace is fine, but the TEI
> has a whole customization mechanism designed to help you constrain and
> extend the TEI as needed. Your TEI ODD customization file should do that
> (or via interfaces like Roma). Do not hand edit the output schemas, they
> are output from the customization file. All uses of TEI are a use of a
> customized version of TEI (even tei_all).  We don't know if you are
> using a particular popular customization, or one you've already made.

Mea culpa: I must admit that I was lazy and avoided to generate the
output schema again for each change. I totally understand the necessity.
Yes, currently I am working with Roma, and I have created a
customization that covers 99% of the things we need.

>> My questions to you, dear TEI-L: Do you see any other options within
>> or outside the TEI? In general, I see this as kind of a quality
>> management layer, which sould not be necessarily a part of TEI, but
>> would be of great interest for other projects. I am sure that I am not
>> the first person that has spent some thoughts on this issue.
>
> If this was general status of the file then @status on <revisionDesc> is
> sometimes used.  (i.e. what milestone is the file as a whole at).  
> Otherwise sometimes people use individual changes inside <revisionDesc>
> to signal some stages in revision, but that is really meant to record
> the revisions which have been made. However, this is often used for
> project workflow purposes.
>
> Since you want to track individual <bibl> why not use @status already on
> this?  As you can see from
> http://www.tei-c.org/release/doc/tei-p5-doc/en/html/ref-bibl.html the
> <bibl> element has the @status attribute available and if your <bibl>
> elements are going through a particular set of stages you can customize
> the TEI (using an ODD or via Roma) to only allow certain values on this.
> 'proofed', 'unchecked', 'updated' or what have you.

That is good to know. So @status seems to be a good idea for <bibl>.
Later on we will require something similar on various descendants of
<msDesc>, e.g. <msIdentifier>, where @status is not yet available. This
will probably default to defining a new attribute. It is probable that
TEI will allow @status also on other elements in the future, or support
any other workflow supporting method on a granular level?

> Hope that helps,
> James

It definitvely does. Thanks!
Torsten


--
Torsten Roeder
Wissenschaftlicher Mitarbeiter
Richard Wagner Schriften

Julius-Maximilians-Universität Würzburg
Institut für Musikforschung
Domerschulstraße 13
D-97070 Würzburg

Tel  +49 (0)931 31-85167
Mail [hidden email]
WWW  http://www.musikwissenschaft.uni-wuerzburg.de/rws
Reply | Threaded
Open this post in threaded view
|

Re: workflow support in TEI?

Stuart A. Yeates
Workflow is fundamentally about document change. At the NZETC we used to use a hand-crafted workflow system based on the <change/> tag. You can see an example in the header of http://nzetc.victoria.ac.nz/tei-source/Cyc05Cycl.xml 

There's no reason why a <change/> couldn't contain a couple of <ref/>s with an @n (indicating a status) and then pointers to a number of elements in that status.

cheers
stuart

--
...let us be heard from red core to black sky

On 18 August 2017 at 00:26, Torsten Roeder <[hidden email]> wrote:
Hi James!

I was wondering to what extent (if at all) TEI supports project workflows. To give an example: Let us assume I want to keep track of all <bibl> elements in <teiHeader> to see what their current status is, regarding the quality of information and the research done so far. Sometimes there is only raw data that was imported or copy/pasted by editors from catalogues which still need to be proofread.

And I assume you want to do this on a per-<bibl> basis. i.e. not just store than they've got up  to bibl123 but that bibl5 has been updated, bibl6 is pending review, bibl7 has been cut-and-pasted only, but bibl8 is entirely done.

Exactly, per-element basis. <bibl> is just one of more examples I could give.

What I have tried so far to approach this task:

* Working with XML comments. This seems manageable when only one or very few editors (=two) are involved. Otherwise it has a good chance to result in chaos.

While that is fine for in-project notes, it isn't a very rigorous system.

* Misusing @cert. This works not that badly, but I am still not happy about it, because @cert refers to the quality of an intervention or an interpretation, and not to the status of editing workflow I intended.

@cert definitely seems wrong.

* Editing my .rnc file, define a project namespace, then define a new @status attribute (something like @my:status), and allow some project specific workflow terms. For now, this is my preferred option.

You do *not* do that by editing your .rnc file. Please, please do not do that. Adding a new attribute, in a new namespace is fine, but the TEI has a whole customization mechanism designed to help you constrain and extend the TEI as needed. Your TEI ODD customization file should do that (or via interfaces like Roma). Do not hand edit the output schemas, they are output from the customization file. All uses of TEI are a use of a customized version of TEI (even tei_all).  We don't know if you are using a particular popular customization, or one you've already made.

Mea culpa: I must admit that I was lazy and avoided to generate the output schema again for each change. I totally understand the necessity. Yes, currently I am working with Roma, and I have created a customization that covers 99% of the things we need.

My questions to you, dear TEI-L: Do you see any other options within or outside the TEI? In general, I see this as kind of a quality management layer, which sould not be necessarily a part of TEI, but would be of great interest for other projects. I am sure that I am not the first person that has spent some thoughts on this issue.

If this was general status of the file then @status on <revisionDesc> is sometimes used.  (i.e. what milestone is the file as a whole at).  Otherwise sometimes people use individual changes inside <revisionDesc> to signal some stages in revision, but that is really meant to record the revisions which have been made. However, this is often used for project workflow purposes.

Since you want to track individual <bibl> why not use @status already on this?  As you can see from http://www.tei-c.org/release/doc/tei-p5-doc/en/html/ref-bibl.html the <bibl> element has the @status attribute available and if your <bibl> elements are going through a particular set of stages you can customize the TEI (using an ODD or via Roma) to only allow certain values on this. 'proofed', 'unchecked', 'updated' or what have you.

That is good to know. So @status seems to be a good idea for <bibl>. Later on we will require something similar on various descendants of <msDesc>, e.g. <msIdentifier>, where @status is not yet available. This will probably default to defining a new attribute. It is probable that TEI will allow @status also on other elements in the future, or support any other workflow supporting method on a granular level?

Hope that helps,
James

It definitvely does. Thanks!

Torsten


--
Torsten Roeder
Wissenschaftlicher Mitarbeiter
Richard Wagner Schriften

Julius-Maximilians-Universität Würzburg
Institut für Musikforschung
Domerschulstraße 13
D-97070 Würzburg

Tel  <a href="tel:%2B49%20%280%29931%2031-85167" value="+499313185167" target="_blank">+49 (0)931 31-85167
Mail [hidden email]
WWW  http://www.musikwissenschaft.uni-wuerzburg.de/rws