This is the mail archive of the xsl-list@mulberrytech.com mailing list .


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

RE: Problem with ignoring system ID of Doctype


At 11:32 2000 09 15 +0100, Paul Terray wrote:
>At 10:09 14/09/00 -0500, Paul Grosso wrote:
>>And all this is currently mostly moot with respect to XML, since the
>>XML spec did not define a catalog (unfortunately, in my opinion).  And
>>since the system id is required, almost all XML processors I know of
>>use it and general ignore the public id.
>
>That's what I have seen till now. And this is a major pain that many tools 
>(usually not parsers, mostly editors) don't know well about back relative 
>entity inclusion, like "../mydir/myent.ent"  (Adept, Stylus being two of 
>them). This means I cannot use the same dtd with Adept (used as a PDF 
>producer) and Omnimark, XSLT (that can use the same, yeah !).

Knowing exactly what is the current "base URI" (against which to 
resolve relative system identifiers) at any time can be tricky, 
but this is something that the W3C XML Activity working groups 
are trying to clarify.  

Meanwhile, I'd be interested to see the example you are considering--in 
exactly what content do you have "../mydir/myent.ent", what are the 
various tools resolving it to, and what do you think it should resolve 
to.  (And if Adept is doing it wrong, I'll work to get that fixed.)

>>However, Java classes that implement catalogs for XML have been put
>>into the public domain, and I'm hopeful that more and more XML
>>processors out there will use them (or the equivalent) to add catalog
>>support for XML.  Norm's article [2] for both a good description of
>>the issues and for a pointer to some Java classes that implement both
>>the TR9401:1997 catalogs and XML Catalogs [3].
>
>I do agree this is useful, but the catalog is not standard yet, and my 
>point here is to unify the DTDs among the tools I use.

I understand.  The best way is really through a catalog (see Norm's
article that I referenced earlier in this thread), but until that is
widely deployed, you can't really rely on public ids.  Now if relative
system ids aren't supported consistently, you can't rely on them either,
so you're stuck with absolute system ids.  If none of these solutions
work, then (sigh) you are stuck, and all I can suggest is to work to
get implementations to support relative system ids properly and/or to
support resolution of public ids.

I'd like to work toward all of the above so users have choices.

paul


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]