Formal Works Referenced: was Re: [osis-core] osisRef (finally)

[Patrick Durusau]

Steve,

Steve DeRose wrote:

<snip>

> That seems fine, but we need different punctuation to separate out the 
> work, so we can tell where the break is. That's in part because 
> low-order fields of work can be omitted. Also, remember we need to 
> settle on a way to find the various bits of a work: auth, title, lang, 
> edition,..... Hey! I just had a thought! What if we had a formal 
> works-referenced sections (in front or back matter, I don't care), 
> which assigned a local name to every work cited. And for each such 
> work, what it does is encode a Dublin Core record! That's got all the 
> right stuff in it. For generic references, you encode a 
> specially-named Dublin Core record that leaves out the irrelevant bits 
> (like edition, language, or date).
>
> That would give us a standardized and relatively complete way of 
> identifying works, and also a short way of referring to them down in 
> the refs; and we wouldn't need any syntax within 'work' since it would 
> be just a name (still would have to delimite with other than '.' if 
> it's optional.
>
> Then all we have to do is define a nice library of predefined names 
> (and make rules about how to extend -- like, un-osis-registered work 
> names have to start with x-).
>
> I think that cleans up a whole mess of stuff. Thoughts?

Well, it would certainly resolve a lot of the problems we are going to 
encounter with the syntax for regex expressions.

Something along the lines of:

<works-referenced>container in header
<work>KJV : required to be valid XML name?
followed by Dublin core elements already declared in the header?

Hmmm, does get us out of the quagmire the syntax for references. The ID 
of work is used by work="" in osisText? Other works can be recorded 
there as well?

Would not take a lot of extension of what we have now to make this work 
fairly robustly.

Still have Harry's problem of declaring a reference syntax. (need to 
look at refDecl but will do that tomorrow)

Troy, Chris, Harry, you will be the early implementers of such texts, 
how does Steve's suggestion sound to you? (Ignore any syntax errors I 
made in interpreting his post, we can fix those.)

Patrick



>
>
>

-- 
Patrick Durusau
Director of Research and Development
Society of Biblical Literature
pdurusau@emory.edu