- From: Norbert H. MIKULA <e_nmiku@utila.ifi.uni-klu.ac.at>
- Date: Sat, 31 May 1997 13:50:48 +0200
- To: w3c-sgml-wg@w3.org
- CC: Tim Bray <tbray@textuality.com>
Tim Bray wrote: > I'm now putting PE reference handling in Lark, and it's turning into > quite a lot of code - since these things are just for DTD's, and Lark > is nonvalidating anyhow, I'm wondering if the extra processing and code > size required are in the spirit of XML, particularly for lightweight > standalone processing. It seems like PE's are largely in the spec to > support authoring-end activities. > > Would it be reasonable to think about saying either that > (a) PErefs should not be used in the internal DTD subset, or > (b) PErefs should not be used in WF docs? As the author of NXP, or in general from the perspective of a software developer, I must say : I write software after a spec. * give me a spec. * PEReferences or not, let's make a decision and stick to it. I can only warn everybody. I believe, if after the next draft, this issue pops up again people will become angry. This issue seems to reappear every once in a while, but I see no progress in making decisions. How many applications of XML as per the draft do we have ? The reason, I believe, is that there is a resistance to approach full conformance because there is not enough certainty. The market is waiting, question is how long it will wait until we end up with one XML interpretation per company. But since I am not only software developer I want to contribute my share to the overall confusion. I would say : PEReferences will be needed for certain applications. Please let's specify how to use them *if* people need it. If they don't need it, fine. If they need it, for instance for maintaining complex DTDs etc., then their users will need more complex software (e.g. a validating XML parsers). Most of the users of NXP, that I have been talking to so far, don't need validation, for instance. Some of them do however. The same goes for PEReferences. (Otherwise I wouldn't have invested the time for implementing them.) Maybe it is a question if we want a "one size fits all" solution or having layers of complexity and conformance. Since we have a distinction between WF and valid, we have already layers, I believe. Thus we could use only the basics for WF and add more complexity for validty. For many (most) XML applications WF will probably be enough. However, I wouldn't like it if I have to move to (costly) SGML just because I want to keep my DTD reasonably maintainable (I am saying this because, I believe, somebody suggested dropping them at all). I personally would say, that we could rule that a parser that needs/wants to check for WF only, doesn't need to process PEReferences. It doesn't really need to process the DTD anyway. -- Best regards, Norbert H. Mikula ===================================================== = SGML, DSSSL, Intra- & Internet, AI, Java ===================================================== = mailto:nmikula@edu.uni-klu.ac.at = http://www.edu.uni-klu.ac.at/~nmikula =====================================================
Received on Saturday, 31 May 1997 07:49:06 UTC