[tei-council] What did we decide?

Sebastian Rahtz sebastian.rahtz at oucs.ox.ac.uk
Mon Jun 4 06:28:46 EDT 2007


Lou Burnard wrote:
>
> 1. TEI Conformance (whether implicit  or algorithmic) requires that 
> there be an ODD. Therefore if you don't have an ODD, you cannot be TEI 
> conformant.
with the caveat that your ODD can be tei_all
>
> 2. Nevertheless, we do create TEI modules as distinct schema/DTD 
> fragments, with published names (someone was actioned to review the 
> names I think? sebastian?) and document said names and modules.
the names are used in an ODD anyway with <moduleRef>, so thats needed 
regardless
>
> So, if you're smart, you can go ahead and build yourself a TEI aschema 
> without an ODD. But we don't promise it will always work -- at release 
> 1.n we might (for example) move an element from one module to another.
and then your ODD will break too.

You can't have it both ways. If we generate DTD and RELAXNG module fragments
and publish them, the punter has a reasonable expectation that a) we won't
willfully break them, and b) we will explain how to use them. Otherwise,
let's stop generating the wretched creatures!

I think you can simply not discuss it, and leave it to IM to document
what actually happens


-- 
Sebastian Rahtz      
Information Manager, Oxford University Computing Services
13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431




More information about the tei-council mailing list