Add publication status via @type in <change>

[annabellasc]

I implemented a feature in our MerMEId of the Telemann Works Catalogue, where we needed a way to mark the publication status of files directly in MEI so they can be processed automatically through pipelines. We plan to use this in a similar way for the Norwegian Heritage Project as well.

The additional MEI structure is part of the revisionDesc and shall be as preceding-sibling of the other change elements:

<change type="false">
        <respStmt>
          <name/>
        </respStmt>
        <changeDesc>
          <p>Partial publication</p>
        </changeDesc>
</change>
<change type="false">
          <respStmt>
            <name/>
          </respStmt>
          <changeDesc>
            <p>Publication</p>
          </changeDesc>
</change>

In <change>, we're using the @type attribute with "true" or "false" to indicate whether a publication (or partial publication) is released or not. In the interface, this is connected to a checkbox. The <changeDesc> contains either "Partial Publication" or "Publication", so the combination of both gives us clear information about the type and release status.

In Telemann we use this In a XSLT script within a git pipeline. Files without type="true" are excluded from the public output, and those marked as "Partial Publication" are filtered to only show basic metadata.

@peterstadler : We talked about this with Arnulf Mattes to use it in the Norwegian Heritage Project, so I'd like to open a pull request to contribute this feature. Feel free to give feedback.

[peterstadler]

The issue was discussed at the MEI slack and while there was IMHO consensus about the need for such a feature, different ways of implementing this have been proposed. E.g.,

• existing MEI features like att.audience or <pubStmt>
• TEI's att.docStatus

The issue will be discussed at the next Metadata-IG on October 13. Maybe @gucl-mu or @KristinaRichts can report back here after the meeting?

[pe-ro]

If a binary published/not published flag is sufficient, then the <unpub> element within <fileDesc>/<pubStmt> is a good solution. The content of the <unpub> element can be restricted by customization for use with MerMEId. When the file is published, the <unpub> element can be replaced by other children of <pubStmt>.

If a historical record of the file is the goal, then using att.docStatus is a better solution.

[ahankinson]

+1 to bringing in att.docStatus, but restricting its use to the status of the MEI document, not the status of the document being described. TEI seems to mix those up. Otherwise "naive" parsers might confuse the two. For example, if an MEI document is given the status of "draft" but the document has the status of "approved" then a parser might think that the MEI document is approved. Keeping the two types of statements separate would be helpful.

The values for TEI generally look OK, but I'm not sure about galley -- this seems to be only really for prints? If it's kept, changing it to something clearer like proof might be helpful?

[gucl-mu]

@ahankinson , thank you for your input! That was exactly what we discussed yesterday in the IG meeting and what we should provide a good definition or explanation for in the guidelines. I think the difference between att.docStatus and <availability> as well as <useRestrict>/ <accessRestrict> needs to be clarified here. That's what it should be about, right?
"Galley" was also a problem for us and is definitely up for discussion.
As Dennis mentioned in the minutes, we plan to revise the first draft at the next meeting and then present it at the ODD meeting before opening a pull request.