[tei-council] MD chapter revised: namespace rules

[Sebastian Rahtz]

Syd Bauman wrote:
> After all, we are talking about two completely different
> use cases. The first, well thought out, documented, and dealt with by
> the creators of P3, is the barrier for document creation, processing,
> and to a lesser extent, interchange. This is the barrier called
> "conformance", which I would like to see set reasonably low.
we do have that fundamental difference in our views :-}
>
>
> Remember that with P3/P4 we set the barrier to conformant
> customization too high, and the result was that most projects used
> TEI Lite, and many hacked it directly. 
>   
that was a technology barrier, not a policy one, I suspect

>
> I disagree. I think there are advantages both for the TEI and for
> users to using the Guidelines for text encoding, not just for
> interchange with the hope of interoperability. For starters, such
> projects will have both less incentive to be TEI-C members, and less
> ammunition when they talk to their deans about becoming TEI-C members. 
> ("OK, so you transform it to TEI to put it in an archive, so what?
> You transform it into HTML for web delivery, do you want us to become
> W3C members, too?")
>   
if they're interested in the TEI, they'll join, even though
they abuse it in their dark inner sanctums.
> The point of much of the syntactic sugar renaming in Tite is to
> reduce keystrokes. Requiring even a 1-character namespace prefix
> would negate any advantages.
>   
yes, I had not grasped that these syntactic sugars
would be in a different namespace until a few hours ago.
I agree, I personally think that's simply unworkable.
>
> Sure. The Imaginary Project creates a schema with half a dozen
> changes vs. vanilla TEI. Rather than expend the effort to figure out
> which things affected by the changes should be in tei: namespace and
> which should be in ip: namespace, they will just put every element
> and attribute affected by the change into the ip: namespace.
>   
I think I get you. You mean that when they constrain
the @type on <div>, they'll put <div> in the ip: space,
because they think they have to? yes, that would be
absurd.
>
> He meant the former, and he obviously does not think it is far too
> late. Nor do I, for that matter. I doubt you really do, either,
> Sebastian, as you understand TEI, but IIRC still have trouble
> wrapping your mind around W3C Schema (although you certainly do
> better than I with it :-)
>   

after a few hours studying an XSD recently, I think
I finally realized how the demented brains
of the designers were working. I still don't
fully understand the ODD processing model,
and I am the one who has to document it :-{

anyway, forget this side meander. The issue is
whether conformance and interchange
are the same thing or not. I conflate them,
you clearly separate them.

we definitely agree that whatever happens,
this is a case where the user community needs
a thorough consultation. And the good news
is that this does not stop us doing other
work in the meanwhile - the definition of conformance
can be revisited again and again (god help us)
over the next 3 or 4 months if
it really has to, I think.
 

-- 
Sebastian Rahtz      

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