Technical Specs

Issues and Suggestions for XML Core Model

Metadata

• Is there a possibility of having an attribute, something like "type", added to <term> so we could further characterize the types of terms? I'm thinking of the traditional triumverate of name, place, and subject, though the list of types could be different. KimTryka?
  • That's a good idea. It will be done unless someone objects. Docs can say that the value of @type should be assigned in view of the @scheme assigned to the parent keywords element, if there is one. TEI allows @type on <term>. WendellPiez
• Also, does it make more sense to associate the attributes "scheme" and "resp" with <term> rather than the element <keywords>? Or else you will need to repeat the <keywords> element if you wanted to use terms from multiple schemes (or if there were multiple people who were assigning the terms? KimTryka?
  • I've made the keywords element repeatable repeatable. I think it's safe to say that keywords can/should be assigned in bunches according to scheme. @scheme is on keywords in TEI. keywords can have a single term if there's only one keyword in a given scheme. Having @resp on term as well as on keywords is a good idea, however. WendellPiez
• Are we going to want a type of term to describe what type of 'thing' the submission is (article, editorial, IM, etc) -- or is that going to be taken care of in some other way (a file naming convention or a convention in the assigned ID for example)? KimTryka?
  • I think you are right that we want an identifier in the file, not a file naming convention, to mark what section a thing falls in; the question is, when that information should be carried authoritatively in the file, and when elsewhere. WendellPiez
  • Do you think we need it in the file earlier than when the PublicationStmt? is added? There is a sort-of general-purpose <assignment> element that could be used for this, as a stopgap. WendellPiez
    • The 'end' that I'm primarily thinking of for this 'means' is resource discovery after publication. In that case, the timing of when the information is added is not the issue, as long as it is there by publication. If adding this information earlier in the cycle aids workflow, all the better. But that depends on how the workflow is to be managed. KimTryka?
    • In one case, the 'this is only there to aid in discovery down the line' case, you might want to do something like:

<term type="dhq-genre">Interactive Media</term>
(or 'section' for 'dhq-genre')

which has the benefit of keeping the information with the rest of theindexing type of information.

• • • In the other case, the 'this information is there to aid some step in the workflow' case, a decision needs to be made as to the appropriate location of the information. Then you might work with Wendell's suggestion of:

<assignment>
   <date value="2005-10-10">Oct 10 2005</date>
   <name role="editor">Kim Tryka</name>
   <item>Edit for Reviews</item>
   <idno type="filename">review-03.xml</idno>
</assignment>

• • • I agree with what's been said so far--important to have information about "section" available for retrieval, and also probably very useful to have as part of workflow tracking. I also think this information is more intrinsic to the file than (say) even the "who reviews it" information, so even if the reviewing assignments got off loaded to the database, I think this information should stay with the file itself. JuliaFlanders
      • probably not a keyword; there's plenty of other information about the document which we'll use for retrieval that won't be treated as keywords, so that's not our only place to look; I think it's more natural to think of this as part of the publicationStmt JuliaFlanders
      • could be duplicated in assignment information, or--maybe better--put there first while the piece is flowing through the workflow. Then, as part of the "publish this now" process, it gets placed in publicationStmt when that is created. A given piece might change its section during the review process, so keeping track of that within while it's still being worked on, and then having the final outcome recorded forever in publicationStmt, would make some sense. JuliaFlanders
      • So--if these intuitions are reasonable, that suggests that we should add a
        element to the bit, and plan as well to include something of this sort in the pubStmt. JuliaFlanders
      • Cool. It will be done. It's nice, easy and small. As to the name 'section' and its definition -- is the distinction btw this and that the latter is open-ended, whereas
        is invalid if not an actual (formal) "section" in DHQ? WendellPiez

-- KimTryka? - 13 Oct 2005