Schematron message 'originalSource' is missing

[juliajbeine]

For "Innocentia victrix" (dracor-org/neolatdracor#80), I get the message that the original source was missing for the digital source. However, the nesting is digital source - digital source - original source. This will be the case for all TransLatin files encoded for NeoLatDraCor. Could the Schematron rule be revised?

       <sourceDesc>
           <bibl type="digitalSource">
              <name>TransLatin</name>
              <idno type="URL"
                 >https://github.com/HuygensING/translatin/blob/main/datasource/transcriptionsOrig/docx/Aler-Innocentia.docx</idno>
              <availability status="free">
                 <p>In the public domain.</p>
              </availability>
              <bibl type="digitalSource">
                 <name>Google Books</name>
                 <idno type="URL">https://play.google.com/books/reader?id=CtBEAAAAcAAJ</idno>
                 <availability status="free">
                    <p>In the public domain.</p>
                 </availability>
                 <bibl type="originalSource"><author>Paul Aler</author>: <title>Innocentia victrix
                       sive Genovefa</title>. <pubPlace>Cologne</pubPlace>: <publisher>Johann
                       Alstorff</publisher>
                    <date>1706</date>. </bibl>
              </bibl>
           </bibl>
      </sourceDesc>

[cmil]

Why do you have two digital sources there? I think in this case the Google Books link should be part of the original source <bibl>.

[lb42]

It wouldn't make sense to add the Google Books link intro the originalSource bibl (I dont believe Google was operating in 1706 !)

[juliajbeine]

As stated in the revised ODD passage on the sourceDesc,
“Please note that digitisations of a print or manuscript published with an identifier and/or under a certain licence cons[t]itute a digital source, e.g. digitisations by the <ref target="https://www.digitale-sammlungen.de/de">Munich DigitiZation Center (MDZ)</ref> or those by <ref target="https://books.google.com/">Google Books</ref>.” (#107)
(As discussed with @lehkost and @peertrilcke.)

I would apply this principle consistently and not start to label GoogleBook links or similar differently when given in a nested bibl element and not at the highest level.

[cmil]

It wouldn't make sense to add the Google Books link intro the originalSource bibl (I dont believe Google was operating in 1706 !)

This is exactly why I would have liked to get rid of that dichotomy of digital vs original source. Without it we could simply nest the <bibl> elements without having to think about those types like that:

<bibl>
  <ref target="https://github.com/HuygensING/translatin/blob/main/datasource/transcriptionsOrig/docx/Aler-Innocentia.docx">TransLatin</ref>
  <availability status="free">
    <p>In the public domain.</p>
  </availability>
  <bibl>
    <ref target="https://play.google.com/books/reader?id=CtBEAAAAcAAJ">Google Books</ref>
    <availability status="free">
      <p>In the public domain.</p>
    </availability>
    <bibl>
      <author>Paul Aler</author>:
      <title>Innocentia victrix sive Genovefa</title>.
      <pubPlace>Cologne</pubPlace>: <publisher>Johann Alstorff</publisher>
      <date>1706</date>.
     </bibl>
   </bibl>
</bibl>

It would mean that the source for the TEI was the Translatin Word document, which in turn was transcribed from the Google Books version which is based on the print from 1706.

If, as in Lacy, you have multiple sources from which the TEI is compiled there could be additional bibl siblings inside sourceDesc.

As stated in the revised ODD passage on the sourceDesc, “Please note that digitisations of a print or manuscript published with an identifier and/or under a certain licence cons[t]itute a digital source, e.g. digitisations by the <ref target="https://www.digitale-sammlungen.de/de">Munich DigitiZation Center (MDZ)</ref> or those by <ref target="https://books.google.com/">Google Books</ref>.” (#107) (As discussed with @lehkost and @peertrilcke.)

I would apply this principle consistently and not start to label GoogleBook links or similar differently when given in a nested bibl element and not at the highest level.

The way I remember the discussion at the Summit Barcamp, the argument for keeping the types digitalSource and originalSource was that we want to have a way to explicitly declare which digital document (if any) was used to create the TEI and which was the (probably printed) source for that document. In that spirit the Google Books instance is not the digitalSource of the TEI but an intermediate.

We don't need the type digitalSource to make clear that a source is of digital nature. Digital sources have a ULR (or ref/@target) that points to a digital document, not to a book on a shelf (or a manuscript in a maze).

I think, if we want to keep the two types digitalSource and originalSource we need a crystal clear explanation of their respective semantics to avoid confusions like the one above.
I would still argue, though, we don't actually need them, the hierarchical nesting of bibl elements already carries all the information we want to encode. Plus, we wouldn't have to decide which type to apply where in all kinds of edge cases.

[juliajbeine]

The way I remember the discussion at the Summit Barcamp, the argument for keeping the types digitalSource and originalSource was that we want to have a way to explicitly declare which digital document (if any) was used to create the TEI and which was the (probably printed) source for that document. In that spirit the Google Books instance is not the digitalSource of the TEI but an intermediate.

As far as I understood the discussion at the DraCor Summit Barcamp, the sourceDesc should include information on some "physical", i.e., printed, source in the inner element, which would be the original source. In most cases, it takes only one or two steps to get to this physical original source (e.g., DraCor TEI - TextGrid - printed book, or DraCor TEI - printed book), but in this case, it is three (DraCor TEI - TransLatin - Google Books - printed book). I do not know how often this is or may be the case throughout the DraCor corpora, but it would be better if the schema covered this, I think.

This is exactly why I would have liked to get rid of that dichotomy of digital vs original source. Without it we could simply nest the elements without having to think about those types like that: […]

I agree with you, the nested <bibl> elements may be sufficient without the type attribute.

[lb42]

I fear we have another ideological difference here. Although the TEI schema does permit bibl to self-nest, all the examples in the Guidelines suggest that the intended semantics of such nesting is that the outer element is contained by the inner one, as when a book chapter or an article contains an "analytic" bibl describing the volume or series in which it appears. See for example the 3rd eg in https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-bibl.html -- this is clearly not the same as the "derived from" relationship you are proposing. For other kinds of relationships between bibls, the TEI proposes a <relatedItem> element (see https://tei-c.org/release/doc/tei-p5-doc/en/html/ref-relatedItem.html)

The new work on Computer Mediated Comms may be relevant here. See for example https://tei-c.org/release/doc/tei-p5-doc/en/html/CMC.html#CMCDocumentSource

Maybe the simplest solution would be to have a flat structure within sourceDesc in which all the related bibls are given at the same level, and relatedItem is used to specify their inter-relationship/s. The @type attribute is useful but not powerful enough to specify all you need to know about a given bibliographic reference -- e.g. its form (digital or analogue), its status in the hiererahcy of transmission of a document, etc.

In Lacy I want to document in the sourceDesc:

• the 'original' source text -- which is usually part of a physical volume, held in some library

• in some cases a note about the manuscript version of the source text, if it exists

• references to the same work in other places (wikidata, viaf, Nicoll, etc.)

• links to digital versions of the same work, some of which may be transcriptions, some of which may be collections of page images

• an indication as to which of these versions has been used as 'copytext', i.e. the version against which the correctness (e.g. pagination, spelling, punctuation) of the TEI transcription has been checked.

[cmil]

@lb42 I don't think the difference is ideological rather than one of the level of TEI expertise. Ever since we started nesting <bibl> elements in DraCor to express the descendance of our immediate source from the original text I assumed that this is a legitimate use of <bibl>. Thanks for pointing out that this is actually not really the intended semantics and also for pointing to <relatedItem> which wasn't on my radar.

After rereading the TEI guidelines following your hints I would now agree that a flat structure could be a viable simple solution. For DraCor we would still need a way to mark what we so far called the "digitalSource" (i.e. the document or resource we used to create the TEI) and the "originalSource" (i.e. the book or manuscript that resource was derived from and from which we would like to to extract the original publication date, place and publisher). If I understand you correctly that's what you call the "copytext" and the "original". I actually like those better than the current DraCor types. In the "Innocentia vitrix" example above it would be clear that the Translatin document is the "copytext" while (in that case) Google Books isn't (it could be in other cases).

So would this then be an acceptable reformulation of the markup (omitting the availability information for demonstration purposes)?

<sourceDesc>
  <bibl type="copytext">
    <ref target="https://github.com/HuygensING/translatin/blob/main/datasource/transcriptionsOrig/docx/Aler-Innocentia.docx">
      TransLatin
    </ref>
    <relatedItem ref="#google-books"/>
  </bibl>
  <bibl xml:id="google-books">
    <ref target="https://play.google.com/books/reader?id=CtBEAAAAcAAJ">Google Books</ref>
    <relatedItem ref="#aler1706"/>
  </bibl>
  <bibl type="original" xml:id="aler1706">
    <author>Paul Aler</author>:
    <title>Innocentia victrix sive Genovefa</title>.
    <pubPlace>Cologne</pubPlace>: <publisher>Johann Alstorff</publisher>
    <date>1706</date>.
  </bibl>
</sourceDesc>  

I'd prefer not to reuse "originalSource" to make a clean cut between the old and new encoding convention. I'm wondering if the <relatedItem> elements would need a @type attribute.

If I understand the TEI guidelines correctly with <relatedItem> nesting would still be an option, so this could be equivalent?

<sourceDesc>
  <bibl type="copytext">
    <ref target="https://github.com/HuygensING/translatin/blob/main/datasource/transcriptionsOrig/docx/Aler-Innocentia.docx">
      TransLatin
    </ref>
    <relatedItem>
      <bibl xml:id="google-books">
        <ref target="https://play.google.com/books/reader?id=CtBEAAAAcAAJ">
          Google Books
        </ref>
      </bibl>
    </relatedItem>
    <relatedItem>
      <bibl type="original">
        <author>Paul Aler</author>:
        <title>Innocentia victrix sive Genovefa</title>.
        <pubPlace>Cologne</pubPlace>: <publisher>Johann Alstorff</publisher>
        <date>1706</date>.
      </bibl>
    </relatedItem>
  </bibl>
</sourceDesc> 

It would have the slight advantage of not needing xml:ids.

[juliajbeine]

I'd prefer not to reuse "originalSource" to make a clean cut between the old and new encoding convention.

I would like to add a comment on “originalSource” and “original”.

As a professor of Shakespeare studies once pointed out, in the case of Shakespeare, one should rather use the term “source text” (“Ausgangstext”) than “original” (“Original(-text)”). He was referring to the Arden edition (third series) of “Julius Caesar” by David Daniell, pp. 121–129. As highlighted there, the First Folio states on its title page that it was “Published according to the True Original Copies” (p. 121). Daniell, thus, discusses what these “True Original Copies” could have been for “Julius Caesar”, an original manuscript (“foul papers”), a further developed version of the manuscript that served as the promptbook for the performance and that the company was working with and on, or a fair copy, with Daniell assuming the third option to be the most likely (pp. 123–129).

Similarly, in the case of Plautus, the extant texts are not the originals written by Plautus. The texts have seen revisions during the performance production process as the troupe has been actively working with the text.

These two cases may illustrate that the term “original” is highly problematic for dramatic texts, especially if they were performed and thus may have been revised in the course of preparing the performance. The textual ‘origin’ of a drama or play cannot always be determined. Thus, I would keep “originalSource” in the sense that this is the original source of the TEI file rather than the original text.

[lb42]

I agree that the word "original" might be considered a bit misleading for plays which have a long history of adaptations. But I think we do need to distinguish between (1) the digital version from which we have derived or against which we have checked the dracor-encoded version and (2) the (usually non digital) version which this digital version claims to represent. I have no problem with using the words "copytext" and "original" for these, as suggested, but am certainly open to other suggestions. How about "copytext" and "source" ?

[cmil]

A type "source" for a source description may seem a bit unspecific or redundant. Would "origin" be less problematic than "original"?

[lb42]

Fine by me

[juliajbeine]

A type "source" for a source description may seem a bit unspecific or redundant. Would "origin" be less problematic than "original"?

Probably, in the sense of “Ausgangspunkt”. I would also check back with @lehkost and @peertrilcke. Did I get it right that the following would be your new suggestion, @cmil?

<sourceDesc>
  <bibl type="copytext">
    <ref target="https://github.com/HuygensING/translatin/blob/main/datasource/transcriptionsOrig/docx/Aler-Innocentia.docx">
      TransLatin
    </ref>
    <relatedItem>
      <bibl xml:id="google-books">
        <ref target="https://play.google.com/books/reader?id=CtBEAAAAcAAJ">
          Google Books
        </ref>
      </bibl>
    </relatedItem>
    <relatedItem>
      <bibl type="origin">
        <author>Paul Aler</author>:
        <title>Innocentia victrix sive Genovefa</title>.
        <pubPlace>Cologne</pubPlace>: <publisher>Johann Alstorff</publisher>
        <date>1706</date>.
      </bibl>
    </relatedItem>
  </bibl>
</sourceDesc>  

I was wondering how to determine a kind of rule for the values of the xml:id attribute of the <bibl> element within the <relatedItem> element, since the TEI files may also indirectly draw on digital versions from the BSB or similar.