Publication statement: idno

[danbalogh]

This is about the <idno> element in the TEI header of documentary (EGD-type) editions.

In #367 we have agreed that the format should be <idno type="filename">DHARMA_INSVengiCalukya00001.xml</idno>, i.e.

• we retain @type="filename" as before
• but add the file extension .xml to the end of the content

Changes have not been made to existing files, and the schema flags an error if I try to put the file extension into this element.
I'd like the PIs (or at least @arlogriffiths ) to confirm that this is indeed how we want to do things.

If yes, I'd like @michaelnmmeyer to implement it:

• update the schema so that the file extension is mandatory in publicationStmt/idno
  • i.e. not forbidden and not simply allowed: the schema should actually check to make sure that it is present
  • I guess this should actually simplify things: the content of <idno type="filename"> must be identical to the name of the file and that's that
• batch insert .xml at the end of all present instances of publicationStmt/idno unless they already end in this string [which is unlikely, as the schema forbids it]

[arlogriffiths]

Changes have not been made to existing files, and the schema flags an error if I try to put the file extension into this element. I'd like the PIs (or at least @arlogriffiths ) to confirm that this is indeed how we want to do things.

I confirm.

[michaelnmmeyer]

<idno> is meant for identifiers. Filenames with the extension .xml are not identifiers in our system. Furthermore, the value of <idno> is not used by the system to assign identifiers in the database (the identifier is derived from the filename instead). It is misleading to the user, who might believe <idno> is used as identifier. <idno> is also redundant, because the filename already provides the information. Thus, I propose we just remove this element. If you absolutely want to keep it, then I suggest we remove the attribute @type="filename".

[danbalogh]

The discussion about <idno> was conducted - and concluded - in the Roles gdoc and in #367 more than a year ago and summarised in #376 in this post in May 2025. The conclusion was as summarised above, and unless one of the PIs now think differently, I think all that remains to do is to implement it. Implementation was actually explicitly requested by Arlo in this post.

That said, if you want to reopen the discussion, I have no preference either way. Anything that you and the PIs can agree on is fine by me. My original preference was <idno type="DHARMA"> and content without the file extension. This is because in my opinion the ID number is a primary and abstract identifier for a DHARMA edition, which happens to be replicated in the filename for convenience, but which can be used in print and in other digital infrastructures too. Anyone who makes a copy of our files to do something with it - e.g. add linguistic or semantic markup, or whatever - can change the actual filename, so the statement that "the filename already provides that information" is only correct in this particular dharmalekha ecosystem.