Dear Mr. Bray, > I believe that XML and SGML have the same handling for CDATA attributes. You are rigth. I was mixing up the SGML-standard and HTML 4. Even though HTML is just an SGML-application, the Browsers are much more liberate in treating markup-errors. When running my docs against nsgmls, href-values like "http://www.test?id=1&id2=2"; produced SGML-parsing errors.
> Lots of people wanted catalogs. Nobody anywhere could introduce evidence > of widely-accepted and interoperable way to handle PUBLIC identifiers, so > they only made it into XML on a private-use basis. If you have your own > catalogs, you can use PUBLIC [check the spec], you just can't try to > interoperate with them, because it doesn't work. If the number of (external or internal) DTD's in an organization grows, the organization WILL have their own catalog-management to get rid of the System-Identifier problems. One major reason is that all XML-Editors I know and several XML-Tools do not support the URI-resolving in system-identifiers. I know this is not your fault! But as you know, many of these tools use their own catalog formats - what leads to ridiculous efforts on the user-side. People WILL have to adjust the XMetal-catalogs, the (slightly different) Adept-Catalogs, the (slightly different) Omnimark-catalogs, the (slightly different) nsgmls-catalogs... There is a REAL need for handling this problems in one of the next versions of XML. I'd suggest that you start by proposing, that only a parser with integrated XMLCatalog (or SGML-Open catalog or whatever format you prefer as long it is only ONE) may be called a conforming parser. > Parser simplification. Unanimous choice of everyone. Lots of these > decisions is the reason that we have dozens of excellent freeware XML > parsers. I was not talking about esoteric SGML features like tag-minimization, inclusions, exclusions or the CONCUR stuff. Small items like the one described do not increase a parsers code in a significant percentage. But still I agree, these decisions had to be made and it is always easy to critizise. > Oh well... I'm probably wasting this lists's bandwidth, since even if > Mr. Pfarr were right, XML is a done deal. -Tim The whole discussion started with a problem that seemed like a parser error to me. I was wrong! I wasted this list's bandwidth! - Armin
