[docbook-apps] Re: [docbook] RE: [docbook-apps] RE: [docbook] RE: [docbook-apps] db 5 and Dita anyone?

2008-05-09 Thread Jirka Kosek
Barton Wright wrote: Another +1 for a non-xinclude conref feature. Well, conref also has its own issues. No one talks about the hidden cost of xincludes, which is that they are expensive in terms of the time they take in a large Java-based doc build process. The more xincludes our doc set

Re: [docbook-apps] Re: [docbook] RE: [docbook-apps] RE: [docbook] RE: [docbook-apps] db 5 and Dita anyone?

2008-05-09 Thread Keith Fahlgren
On Fri, May 9, 2008 at 1:53 PM, Jirka Kosek [EMAIL PROTECTED] wrote: There is no reason why XInclude should be slower then conref, actually I would expect that conref will be little bit slower as it has to be implemented during XSLT processing, not just during XML parsing. But you have to

Re: [docbook-apps] Re: [docbook] RE: [docbook-apps] RE: [docbook] RE: [docbook-apps] db 5 and Dita anyone?

2008-05-09 Thread Jirka Kosek
Keith Fahlgren wrote: That said, we've experienced slowness with deep XIncludes on all of our systems (lots of architectures, OSs, tools). I do not believe their is anything broken in our toolchains. Do you reference DTDs from XIncluded files? If so, are DTDs fetched from catalog? This could