[osis-core] ID confusion

[Patrick Durusau]

Greetings,

Troy called last night (after my bedtime but got the voice mail) to 
discuss the question of self-identification and IDs. In writing a reply 
this morning, I have "discovered" an ambiguity that you may all have 
been aware of but is new to me.

We have ID attributes on all elements (well, verseID on verses but more 
on that in a moment).

If IDs or verseIDs on verses serve as canonical identifiers (Matt.1.1, 
Job.5.5) then they have a different semantics from the IDs on pointers, 
which are just IDs in the traditional sense of SGML/XML IDs, in that 
they only identify the element for the document instance. They have no 
meaning outside of the document.

When I was using "Who am I?" I meant it in terms of identification in 
the traditional SGML/XML sense and not in the sense of a canonical 
identifier.

For economy reasons I don't have a problem with verseID or other IDs 
operating as canonical identifiers but I think we will need to document 
the difference in the semantics. Already part of the way there with 
verseID, which is not a true SGML/XML ID, although it has some of the 
characteristics of one.

So, for other texts, Josephus for example, do we need to allow other 
elements to have something like verseID that will allow them to have a 
canonical reference without overloading ID or should we simply overload 
ID and document it?

What brought this into focus for me was writing out that pointers have 
IDs that are not canonical references, just the ID of the pointer in the 
document but that they point to IDs that are canonical references (as 
well as being part of a document).

More coming on the reference syntax.

Patrick

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