At 07:22 PM 12/29/2006, Sebastian Rahtz wrote:
>>I think you've allowed something very important here. Tying
>>conformance to the ODD may enable us to track the changes more
>>easily, but smart guys like Ron are still plenty capable of working
>>up methods of validating documents to TEI subsets which, though
>>they may not so easily be "provable" as TEI, are still nonetheless
>>TEI in any sense that means anything.
>maybe. but I claim the ball is in your court to define how we will
>test that conformance in real life.
As I've said before, I thought the P3/P4 definition had the balance
just about right. It pointed to validation against a DTD and defined
ways in which a custom DTD could vary while remaining in conformance,
while not relying on exclusively automated mechanisms. It was not a
cookie-cutter approach, but instead held projects to a high standard.
Conformance meant not simply formal validation as such, but
intellectual rigor, demonstrated through documentation and explicit
rationales that showed how and why modeling decisions were made and applied.
>>They may even have to, from time to time, due to unavoidable
>>limitations in Roma or the TEI architecture.
>again, maybe. but do these limitations exist now?
>>But up to the moment he pushes that button, he's not conformant?
>things are conformant when they offer themselves to public
>scrutiny and the peer group agrees that they are so. just like
>in the law, where you have committed the crime when the judge
>and jury say you have. There is no external force saying absolutely
>"Wendell, thou art guilty". With apologies to the religious folks
>amongst you, but to my mind all laws are man-made.
Ah, so you agree with me then. So, if a project has no ODD but can
justify their means and methods, they can yet be blessed? Will the
peer group be willing to acknowledge that ODD implementations are no
better as final arbiters than the schemas they define, and that
indeed there might be a category (albeit narrow) of TEI that may (now
or in future) have good reasons to use other mechanisms?
>> any meta-process excludes edge cases which may then prove to be
>surely then we re-examine the meta-process and improve it, not
>throw it away?
Elaborating the meta-process is a "solution" that often undermines
the very reason for having the meta-process in the first place.
(And FWIW, no one has suggested throwing it away. When a meta-process
fails, one can return to process, but this doesn't mean the
meta-process can't be useful otherwise.)
>I er confess that I have lost track of what thing Ron found which started
>all this? was in just "how do I remove numbered divs"? if so, I firmly
>believe that is simply a bug which can and will be fixed.
This is highly relevant. Ron demonstrated that he could make a legal
modification that would not pass conformant XML DTD parsers due to
ambiguous content models. Lou argued that this problem comes with the
territory, implying that there is really no good fix. Ron can edit
the DTD by hand and get what he wants (while eschewing the ODD/Roma),
and indeed produces documents that would be valid to another
Roma-generated DTD (one more permissive than he wants). But he has no
ODD mechanism for defining the correct schema. (Ron, please feel free
to leap in if my paraphrase is not correct.)
One other thing occurs to me: what happens if Ron uses RNG? It would
be nice if the problem would then just go away (though no help to DTD users).
Cheers and best wishes for 2007,
Wendell Piez mailto:[log in to unmask]
Mulberry Technologies, Inc. http://www.mulberrytech.com
17 West Jefferson Street Direct Phone: 301/315-9635
Suite 207 Phone: 301/315-9631
Rockville, MD 20850 Fax: 301/315-8285
Mulberry Technologies: A Consultancy Specializing in SGML and XML