How may we improve the encoding of the source(s) of a DraCor TEI file in the sourceDesc?

[juliajbeine]

Overview

As discussed during the DraCor Summit and some GitHub issues (#107, #161), we may revise the encoding of the source(s) of a DraCor TEI file in the <sourceDesc> element. The aim is to better define and clarify the different source types and to cover the various ways that DraCor TEI files are created through the DraCor corpora. For instance, while most of the DraCor TEI files have been created from existing digital sources, some are now created from scratch.

We may exchange ideas in this GitHub discussion and later document the discussion results in the related issues rather than have lengthy conversations in different GitHub issues.

Generally, in our previous discussions, we have seen two areas that need improvement:

1. Revise and define the source types given in the <bibl> elements via the type attribute. They should cover all scenarios of TEI file creation for DraCor.
2. Structuring the elements in a way that matches the TEI P5 guidelines, most probably by dissolving the current nested structure (see this comment).

We may proceed as follows:

1. Create an overview of the scenarios of how TEI files are created for DraCor.
2. Apply the current source type terminology of originalSource and digitalSource to all these scenarios and assess the current usage of the terminology throughout the DraCor corpora exemplarily.
3. Present the alternative source type terminology currently under discussion (“copy-text” and “origin”) and apply it to the scenarios.
4. Clarify what aspects of the TEI file creation the new source type terminology should cover.
5. Present a new source type terminology, probably by revising the previous two terminologies (see 2. and 3.).
6. Apply the new terminology to all scenarios.
7. Assess the current, discussed, and newly proposed terminologies in comparison.
8. Discuss how to structure/nest/relate the elements within the <sourceDesc> element.

[juliajbeine]

1. Scenarios of TEI file creation for DraCor

Scenario A

The DraCor TEI file is based on a digital text (such as provided by Project Gutenberg, TextGrid, or CENETON) based on a printed edition.
Example(s): Gottsched’s “Der sterbende Cato” (GerDraCor)

1. DraCor TEI file
2. TextGrid text file published online
3. printed text edition that is the basis of the TextGrid file

Note: Currently, this is the most common scenario throughout the DraCor corpora.

Scenario B

The DraCor TEI file is based on a digital scan of a printed edition. The book scan has been published and licensed online (e.g. by a national library or Google Books).
Example(s): Achat’s “Ein April-Scherz” (GerDraCor)

1. DraCor TEI file
2. Google Books scan that is the basis of the TEI file
3. printed text edition that was digitised by Google Books

Note: This scenario seems to be one of the more frequent ones in DraCor.

Scenario C

The DraCor TEI file is based on a digital text, based on a book scan which has been published and licensed online and is based on a printed edition.
Example(s): Aler’s “Innocentia victrix” (NeoLatDraCor):

1. DraCor TEI file
2. TransLatin transcription published online
3. Google Books scan that is the basis of the transcription
4. printed text edition that was digitised by Google Books

Note: This scenario pertains to more than 100 dramatic texts from the TransLatin project at least.

Scenario D

The DraCor TEI file is based on a manuscript without an intermediary (digital) source.
Example(s): “Theodosius Iunior Princeps Sapiens, ac Felix” (Ljubljana Jesuit Drama, NeoLatDraCor)

1. DraCor TEI file (being prepared)
2. manuscript in the OeStA/HHStA archive that is transcribed and then encoded for NeoLatDraCor

Note: This scenario seems not to have been very present in the TEI file creation until now. However, we will have ca. 40 dramatic texts added to NeoLatDraCor based on manuscripts without an intermediary (critical) text edition and without an intermediary digital source starting from 2026 via the Neo-Latin School Drama and Ljubljana Jesuit Drama projects.

Scenario E

The DraCor TEI file is based on a born-digital text.
Example(s): I could not find any DraCor examples spontaneously.
Hypothetical example:

1. DraCor TEI file
2. born-digital text

Scenario F

The DraCor TEI file is based on a printed edition without an intermediary (digital) source.
For illustration, the creator of the TEI file may have transcribed the text from the book manually or they may have scanned and OCRed the text before encoding it.
Example(s): I could not find any DraCor examples spontaneously.
Hypothetical example:

1. DraCor TEI file
2. printed text edition

(Please let me know if you would like to revise a scenario or add another one.)

[juliajbeine]

2. Current source type terminology

Terminology

<bibl type="digitalSource"> refers to a digital surrogate of a ‘physical’ print or manuscript source. A digital source may be a born-digital (critical) edition or a digital (critical) text edition based on a print or manuscript source. Digitisations of a print or manuscript published with an identifier and/or under a certain licence constitute a digital source.

<bibl type="originalSource"> refers to a print or manuscript source. The original source may be the basis of the digital source of the TEI file or the direct source of the TEI file if the TEI file was created from scratch.

(For more details, see the DraCorODD or the updated version for better clarification.)

Observations on current usage

• Some DraCor TEI files, e.g. this file in CalDraCor, link to a digital surrogate of a printed text edition within the <bibl type="originalSource"> which does not match the definitions of “digital source” and “original source”.
• In SpanDraCor, the encoders have introduced their own type terminology that mixes information usually found in <standOff> and <sourceDesc> in DraCor TEI files (e.g. in this file).
• In ShakeDraCor, the encoders give their digital text source and the print source of that digital text in the <sourceDesc>, but without any <bibl type> elements. They encode all the information in <biblFull>.
• In SweDraCor, the encoders give their digital text source in the <publicationStmt> and the print source of that digital text in the <sourceDesc>, but without any <bibl type> elements. Moreover, they evaluate the CC-BY-SA licence as restricted and not free, although this licence is generally classified as “free” (see this article), so maybe we should define what the values “free” and “restricted” mean in the availability statement.
• At least in SweDraCor, ShakeDraCor, and ItaDraCor, DraCor seems not to be given as the publisher in the <publicationStmt>, but the previous project that serves as a source for the DraCor corpus is. Maybe this explains the individual encoding of sources in SweDraCor.

Scenarios

Scenario A

Example(s): Gottsched’s “Der sterbende Cato” (GerDraCor)

1. DraCor TEI file
2. TextGrid text file published online = <bibl type="digitalSource">
3. printed text edition that is the basis of the TextGrid file = <bibl type="originalSource">

Scenario B

The DraCor TEI file is based on a digital scan of a printed edition. The book scan has been published and licensed online (e.g. by a national library or Google Books).
Example(s): Achat’s “Ein April-Scherz” (GerDraCor)

1. DraCor TEI file
2. Google Books scan that is the basis of the TEI file = <bibl type="digitalSource">
3. printed text edition that was digitised by Google Books = <bibl type="originalSource">

Observation: Some DraCor TEI files, e.g. this one, link to a digital surrogate of a printed text edition within the <bibl type="originalSource"> which does not match the definitions of “digital source” and “original source”.

Scenario C

The DraCor TEI file is based on a digital text, based on a book scan which has been published and licensed online and is based on a printed edition.
Example(s): Aler’s “Innocentia victrix” (NeoLatDraCor):

1. DraCor TEI file
2. TransLatin transcription published online = <bibl type="digitalSource">
3. Google Books scan that was the basis for the transcription = <bibl type="digitalSource">
4. printed text edition that was digitised by Google Books = <bibl type="originalSource">

Problem: The current schema does not allow more than one <bibl type="digitalSource"> (see this issue).

Scenario D

The DraCor TEI file is based on a manuscript without an intermediary (digital) source.
Example(s): “Theodosius Iunior Princeps Sapiens, ac Felix” (Ljubljana Jesuit Drama, NeoLatDraCor)

1. DraCor TEI file
2. manuscript in the OeStA/HHStA archive that has been transcribed and then encoded for NeoLatDraCor = <bibl type="originalSource">

Scenario E

The DraCor TEI file is based on a born-digital text.
Hypothetical example(s):

1. DraCor TEI file
2. born-digital text = <bibl type="digitalSource">

Scenario F

The DraCor TEI file is based on a printed edition without an intermediary (digital) source.
For illustration, the creator of the TEI file may have transcribed the text from the book manually or they may have scanned and OCRed the text before encoding it.
Hypothetical example:

1. DraCor TEI file
2. printed text edition = <bibl type="originalSource">

[juliajbeine]

3. Alternative source type terminology A (under discussion)

Background

Based on the discussion with @lb42, we may consider using the terms “copy-text” and “origin”.

Considerations

• In English-language philology, the term “copy-text” refers to the direct source of a scholarly text edition, without including a digital or non-digital format as a condition in the definition.
  For instance, the Oxford Reference defines “copy-text” as follows: “The specific text used as the basis for a later edition of a given work. The scholarly editor of a literary work by a deceased author will decide upon the copy-text and reproduce this, accompanied by lists of variant readings found in other editions (or manuscripts) of the same work. Standard editorial procedure is to adopt as the copy-text the last edition of the work that was published during the author's lifetime; but there may be strong reasons for preferring the first published edition, or a manuscript version, or a set of proofs corrected by the author.”
  Thus, this term would not be suitable to assess the format of the related sources (digital or non-digital). Moreover, the term seems to be closely related to critical editions, which might be misleading in the context of DraCor.
  If we went with this term, we may point out that it is used in the context of critical editions, but that DraCor primarily uses it to mark the direct source of the DraCor TEI file, without DraCor providing critical editions.
• The term “origin” seems a little bit less problematic than the term “original”, especially considering the difficult and rather textual transmission history of some dramatic texts in DraCor (see this comment and the succeeding comments).

Tentative definitions of the alternative source types

<bibl type="copy-text"> refers always to the direct source of the DraCor TEI file, may it be digital or non-digital.

If the DraCor TEI file refers to a digital source based on a ‘physical’ source, i. e. a printed text edition or manuscript, the ‘physical’ origin is given by <bibl type="origin">.

Scenarios

Scenario A

Example(s): Gottsched’s “Der sterbende Cato” (GerDraCor)

1. DraCor TEI file
2. TextGrid text file published online = <bibl type="copy-text">
3. printed text edition that is the basis of the TextGrid file = <bibl type="origin">

Scenario B

The DraCor TEI file is based on a digital scan of a printed edition. The book scan has been published and licensed online (e.g. by a national library or Google Books).
Example(s): Achat’s “Ein April-Scherz” (GerDraCor)

1. DraCor TEI file
2. Google Books scan that is the basis of the TEI file = <bibl type="copy-text">
3. printed text edition that was digitised by Google Books = <bibl type="origin">

Scenario C

The DraCor TEI file is based on a digital text, based on a book scan which has been published and licensed online and is based on a printed edition.
Example(s): Aler’s “Innocentia victrix” (NeoLatDraCor):

1. DraCor TEI file
2. TransLatin transcription published online = <bibl type="copy-text">
3. Google Books scan that was the basis for the transcription = <relatedItem>
4. printed text edition that was digitised by Google Books = <bibl type="origin">

Comment: The source type terminology of “copy-text” and “origin” would not cover the intermediary Google Books scan. But we may use < relatedItem> to encode it (see this comment by @lb42).

Scenario D

The DraCor TEI file is based on a manuscript without an intermediary (digital) source.
Example(s): “Theodosius Iunior Princeps Sapiens, ac Felix” (Ljubljana Jesuit Drama, NeoLatDraCor)

1. DraCor TEI file
2. manuscript in the OeStA/HHStA archive that has been transcribed and then encoded for NeoLatDraCor = <bibl type="copy-text">

Comment: According to the tentative definitions, the manuscript would be classified as “copy-text”, and the type “origin” would not be relevant here as there is no intermediary digital source.

Scenario E

The DraCor TEI file is based on a born-digital text.
Hypothetical example(s):

1. DraCor TEI file
2. born-digital text = <bibl type="copy-text">

Comment: According to the tentative definitions, the manuscript would be classified as “copy-text”, and the type “origin” would not be relevant here as there is no intermediary digital source.

Scenario F

The DraCor TEI file is based on a printed edition without an intermediary (digital) source.
For illustration, the creator of the TEI file may have transcribed the text from the book manually or they may have scanned and OCRed the text before encoding it.
Hypothetical example:

1. DraCor TEI file
2. printed text edition = <bibl type="copy-text">

Comment: According to the tentative definitions, the printed text edition would be classified as “copy-text”, and the type “origin” would not be relevant here as there is no intermediary digital source.

[juliajbeine]

4. Aim and scope of the DraCor source type terminology

Against the backdrop of the current and discussed alternative source type terminologies, we should consider the following questions before implementing new source type terminologies from my point of view.

1. To what extent would we like to document the sources of the DraCor TEI files?
   For instance, the current source type terminology has aimed at tracing the ‘source dependencies’ until it ‘reaches’ a ‘physical’ source, i.e. the printed book one may hold in one’s hands:
   physical text (- intermediary (digital) source) - DraCor TEI file.
2. Should we only document text versions, editions, etc. related to the process of creating the respective DraCor TEI file or include other texts (see the observation about SpanDraCor above)?
3. Would we like to assess the format of the sources (digital – non-digital)?

@lehkost, @peertrilcke, @cmil, what do you think?