[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [dita] Fwd: DITA grammars
Rodolfo is not correct. RNG provides for attribute defaulting (DITA depends on it) and I believe George Bina implemented support for it in the general Java RNG processing tools some years ago now. As for not supporting XSD, another important issue and reason for not supporting it is that XSD 1.0 *is fundamentally broken* as specified with regard to the redefine feature, on which DITA XSDs as defined for DITA 1.x depend. That is, using the same XSDs you may get different, an incompatible, results from different *conforming* XSD processors. It happens that Xerces behaves the way DITA needs but others do not (I think the .NET processor does not but I canât swear to it). XSD 1.1 added the new override feature, which fixes the issue with redefine and provides a much better match to DITAâs requirements, but I donât think there are many XSD 1.1 implementations out there. I think Krisâ responses are the correct ones: RNG is normative, DTD is maintained by the TC, and XSDs are generatable. That said, I think itâs reasonable for the DITA community to maintain essentially the same XSDs we did for 1.x, which can be generated using the existing (or updated) rng2xsd transform I developed for DITA 1.3. Cheers, E. -- Eliot Kimber http://contrext.com From: <dita@lists.oasis-open.org> on behalf of Kristen James Eberlein <kris@eberleinconsulting.com> Feedback from Rodolfo Raya ... Best,
Hello Kris, Moving DITA grammars to RelaxNG looks cool and a modern thing to do, but from my perspective as developer this was a bad idea. When a Java-based parser reads an XML file that declares a DTD or XSD grammar, the parser sets element attributes to their default values when the user doesnât explicitly set in the XML file. If the parser reads a DITA file, it automatically sets the @class attribute from the grammar, even if the author skipped it (most common case). Same with any attribute that has a default in the grammar. Reading an XML file that declares a RelaxNG grammar using a Java-based parser does not set skipped attributes to their default values. Reading a DITA file with RelaxNG grammar does not let you know the value of the @class attribute in its elements. My translation tools depend on the @class attribute to process user-defined elements. This is a problem when users try to adopt RelaxNG for their customized grammars. At least my tools canât support them today. Itâs not a new thing, the problem appeared long ago. So, the move to embrace RelaxNG creates problems for some tool vendors. It requires major changes in XML parsers and catalog managers. I really donât know if the DITA TC considered the libraries that developers use on different platforms, like .NET on Windows or Java. My impression is that the TC looked at authoring and publishing environments only. Take this email as a small contribution to the discussion on grammars. Iâm not requesting any change or reconsideration, just providing additional information. Regards, Rodolfo -- --------------------------------------------------------------------- To unsubscribe from this mail list, you must leave the OASIS TC that generates this mail. Follow this link to all your TCs in OASIS at: https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php |
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]