[osis-core] begin/end [was: When an osisID is not an osisID?]

[Patrick Durusau]

Troy,

Agree we did try to avoid some of the more difficult implementation 
issues to get a schema out and failed. I do think we can finish the 
current version and get it out without too much more delay.

Thanks for remembering what the problem case was for the ugly syntax!

Other than just arbitrary convention, is there some reason why I could 
not do the following:

(Also uses present attribute names)

<verse osisID="John.1.1;John.1.2;John.1.3">, etc., so that a simple 
substring match will find that John.1.2 is contained here?

Perhaps just for my own clarity: The target (of any pointer but in our 
case of cite, outCite) is always osisID. The pointer, is always cite or 
outCite, which can address a range of osisIDs.

The reason I point that out now is that I don't see the same 
grainularity requirement in cite/outCite as I do for Troy's example of 
osisID. For pointing to be successful by simple software/mechanisms, we 
need to avoid complexity of osisID. On the other hand, so long as 
resolving a reference results in you being delivered to the first 
matching osisID, I think that is good enough for this release.

Does that help?

Patrick

Troy A. Griffitts wrote:

> Ok, after rethinking things with the begin/end, I'm starting to 
> remember why we were requiring the ugly markup for these case as 
> opposed to the much nicer markup (both stated below):
>
>
> ugly:
>
> <verseStart refID="John.1.1" />
> <verseStart refID="John.1.2" />
> <verseStart refID="John.1.3" />
> (John 1:1-3) In the beginning...
> <verseEnd refID="John.1.1" />
> <verseEnd refID="John.1.2" />
> <verseEnd refID="John.1.3" />
>
>
> clean:
>
> <verse beginOsisID="John.1.1" endOsisID="John.1.3">
> <reference osisRef="John.1.1-John.1.3">John 1:1-3</reference> In the 
> beginning...
> </verse>
>
>
> The ugly markup make for much easier code.  Consider this reference:
>
> <reference osisRef="John.1.2">John 1:2</reference>
>
> In the ugly code, we need no knowledge of reference schemas.
> Introducing begin/end introduces the need for a hugely complex 
> mechanism to know that the 'clean' code, above, is a valid target for 
> this reference.
>
> If my memory serves me, we decided, in Dallas, to use the ugly coding 
> so we might put off the 'hugely complex mechanism', for now.
>

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