#74 Add basic functionality for INTERNAL apparatus to XSLT

[BODARD Gabriel]

If someone would be willing to create a basic template for generating a below-the-text apparatus division from common features such as reg/orig, sic/corr, app/rdg in the edition div. This should be (a) much simpler than the current $app-style="ddbdp" template in there, and (b) use an $app-style such as "internal" aut sim.

Any volunteers to work on this?

[BODARD Gabriel]

A mini version of this now exists (only takes account of choice/reg/corr, not app/lem/reg or anything else interesting) with the parameter $apparatus-style='iospe'. Something better could (and should) be done, so this ticket remains open and seeking a volunteer...

[Tom Elliott]

I'm taking ownership of this ticket in order to organize an implementation sprint in october 2014.

[Tom Elliott]

Now shooting for implementation sprint in November 2014.

[Tom Elliott]

Reassigned to Pietro for first pass at implementation, per agreement in sprint meeting 3 November 2014.

[Pietro Maria Liuzzo]

a version of the changes to address this issue according to the discussed architecture is under development and includes
- break down in small templates of each case
- a new value "defaultmax" for internal-app-style parameter
- test cases of functionality for both changes according to leiden-style and changes according to apparatus parameter

[Pietro Maria Liuzzo]

Meeting for feedback and further development proposed for 20/1/15

[Pietro Maria Liuzzo]

If you take this sample text randomly picked from IGCyr and modified to contain a several of the features we want to support in internal apparatus, and try to transform it with the example-xslt, the following results will be given.

  <div type="edition" xml:lang="grc" xml:space="preserve">
  <div type="textpart" subtype="column">              <ab>
                    <lb n="1" rend="left-to-right"/><w part="I">τ<corr>ι</corr>μο</w><gap reason="lost" unit="character" extent="unknown"/><!--Correction Without Specification-->
                    <lb n="2" rend="right-to-left"/><gap reason="lost" unit="character" extent="unknown"/><w part="F">εν</w> <g type="tripunct">⋮</g> τοι <w lemma="πρόγονος">πρ<lb n="3" break="no" rend="left-to-right"/>ογονο<unclear>ι</unclear></w> <gap reason="lost" unit="character" extent="unknown"/><w>α<supplied reason="lost">ὐ</supplied>θ<supplied reason="lost">έν</supplied>τ<choice><corr>ο</corr><sic>α</sic></choice>υ</w> <!--Erroneous Substitution-->
                    <lb n="4" rend="right-to-left"/><choice>
 <reg>τιμὴν</reg>
 <orig>τειμην</orig>
</choice><!--regularization--><gap reason="lost" unit="character" extent="unknown"/><w part="F">ι</w> <g type="tripunct">⋮</g> τοι <persName type="attested"><name nymRef="Πρόκλος">Πρό<lb n="5" break="no" rend="left-to-right"/><choice>
 <unclear>Γ</unclear>
 <unclear>Ε</unclear>
 <unclear>Ζ</unclear>
 <unclear>Ξ</unclear>
 <unclear>Π</unclear>
 <unclear>Σ</unclear>
 <unclear>Τ</unclear>
</choice><unclear>ϙ</unclear>λō</name></persName> <g type="dipunct">:</g> <w part="I">θυγ<supplied reason="lost">ατ</supplied></w><gap reason="lost" unit="character" extent="unknown"/><!--Ambiguous Characters With Alternatives Offered --></ab></div>
               <div type="textpart" subtype="column"><ab>     <lb n="6"/>  <subst>
 <add place="inline">τέσσαρας</add>
 <del rend="corrected">τεσσαρες</del>
</subst> <!--Ancient Corrections (Where Both Old And New Text Can Be Read)-->

<lb n="7"/> public<add place="overstrike">a</add>e  <!--Ancient Corrections (Old Text Lost)-->
<lb n="8"/> <add place="above">αβ</add> <!--Addition By Ancient Hand -->
<lb n="9"/> <expan>πρ<hi rend="superscript">οσ</hi>
 <ex>γίνεται</ex>
</expan>
<!--Raised/Lowered Characters-->
<lb n="10"/> <app type="editorial">
 <lem source="#BL_3.8">ἡ<supplied reason="lost">
   <expan>προγ<ex>εγραμμένη</ex>
   </expan>
  </supplied>
 </lem>
 <rdg>ὡ<supplied reason="lost">ς <expan>
    <ex>ἐτῶν</ex>
   </expan> ιϛ</supplied>
 </rdg>
</app> <!--editorial--> 
   <lb n="11"/> <app type="alternative">
 <lem>
  <expan>Λευκ<ex>ο</ex>γ<ex>ίου</ex>
  </expan>
 </lem>
 <rdg>Λευκῶ<add place="above">ν</add>
 </rdg>
</app> <!--alternative-->
</ab></div>
            </div>

internal-app-style = 'default'
NOTHING AT ALL

internal-app-style = 'iospe'
3: orig. Α.Θ..ΤΑΥ; 6: corr. ex ; 9: οσ s.l.

internal-app-style = 'minex'
l.3 sic, orig. α | l.4 reg., orig. τειμην | l.10 #BL_3.8; ὡς ἐτῶν ιϛ | l.11 Λευκῶν

internal-app-style = 'fullex'
l.1 ι corr. | l.3 sic, orig. α | l.4 reg., orig. τειμην | l.5 Γ aut Ε aut Ζ aut Ξ aut Π aut Σ aut Τ | l.6 del τεσσαρες | l.7 a del | l.8 αβ add above | l.9 superscript | l.10 #BL_3.8; ὡς ἐτῶν ιϛ | l.11 Λευκῶν

The problems which I still have are:

• for some reasons I obtain nothing with xsl:apply-templates select="t:rdg" I do not understand why. I have put value-of for the time being, but l.10 should be ὡ[ς (ἐτῶν) ιϛ], I think
• why are the variant unclear in the choice like that in the text?
• what should I do with @source in lem?

[Tom Elliott]

I am taking ownership of this ticket with the intention of scheduling a meeting soon to read on all critical parties (DLL, DCLP, etc.) with view to getting this resolved.

[BODARD Gabriel]

Revisiting this ticket: I think we had (begun to) implement Pietro's example above, with a minimal example (internal-app-style = 'minex') based mostly on IOSPE, and a fuller example (internal-app-style = 'fullex') based at least partly on DDBDP. Tom will check how this is implemented currently, and also compare with the needs of the literary projects, and convene a sprint on the question if needed.

[BODARD Gabriel]

One issue that has arisen in training and elsewhere is that people are confused about the distinction between "internal" (parallel) and "external" (reference) apparatus, exacerbated by the fact that we don't have robust XSLT transformation to demonstrate the output of either.

[Hugh A. Cayless]

Notes re internal and external apparatus:

Emmanuelle makes the excellent point that there's a cost/benefit tradeoff for projects deciding to do full parallel segmentation apparatus vs. simple notes on the text. For DLL, the carrot is that users get to play with the reading text using the apparatus. Gabby also points out that inline apparatus can be very useful for downstream applications that wish to do text processing (if they want to treat particular types of substitution, for example). There's also a consideration with internal vs. external app regarding what goes into the apparatus. If things like corrections and normalizations or ancient diacritics feature in the apparatus when the document is displayed in a web or print version, then an internal apparatus is probably a requirement (otherwise you're inteleaving internal and external features).

A simplified apparatus best practice is a desideratum. The DDbDP approach is both dated and overly complex and so should be deprecated.

[Tom Elliott]

I have sent doodle poll invitations for this meeting to all who have participated in this thread with the goal of having the meeting within the next 4 weeks. Please email me or append here if you have not received the doodle poll invitation but would like to participate.

[Pietro Maria Liuzzo]

https://sourceforge.net/p/epidoc/code/2760/ is the edit to the guidelines, pending approval of https://github.com/EpiDoc/Stylesheets/pull/5 which deals with issues described in the agenda of the meeting of 29.01.2020

[Tom Elliott]

The long-promised meeting was held on 29 January 2020. In attendance: @hcayless @pietroliuzzo @gabrielbodard and myself. We identified the following major tasks:

1. Collation: Do a proper survey and concordance of what is handled by the stylesheets now, which parameter(s) govern associated behavior, and whether they work as expected.
2. Run that survey past some section of constituency and see if they think it's good enough for core (markup? EDAG?)
3. If 1&2 reveals changes that need to be made to the stylesheets, implement them.
4. Assess completeness, accuracy, and fitness for purpose of the Guidelines in light of results of the above.

@pietroliuzzo volunteered to make the first pass at the collation task. He has shared the results with meeting participants and has also followed this up by suggesting a number of changes to the stylesheets and the Guidelines as reflected in his comment 3 days ago (see above).

We now need someone(s) to review the stylesheet and Guidelines changes in light of the collation.

[Hugh A. Cayless]

Tom and I have been reviewing Pietro's PR and think at present it may need some modifications. Action on us to resume looking at it and get back to Pietro.

[BODARD Gabriel]

@hcayless @paregorios do you expect this to be implemented and the pull request accepted by September 2020, or should we punt this?