On Sat, May 04, 2002 at 10:20:25PM +0200, Daniel Egger wrote: > Am Sam, 2002-05-04 um 20.42 schrieb Ayose: > > It should be DockBook/XML :) > > Well, is IS XML, just the DTD is an SGML one. All strict XML like > correct closing and shortened tags as well as case sensitivity are > obeyed, I just haven't checked if anything special changed between > DocBook/SGML and DocBook/XML but it should be trivial to fix that. Assuming you are using the version 4.1.2 DTD, there is no difference. If you are using the older version 3.x DTD, then to move to 4.1.2 XML the only significant change is that whatever the tag was called that gave the article info (title, authors, etc) is now called <articleinfo> (but was called something else in 3.x -- sorry, can't remember the old name). > > Yes, I have seen it, but I think that it is better XSLT instead of > > python, because XSLT is more easy and it was designed for this kind of > > jobs :-). > > Well, DSSSL wasn't sufficient layout-wise and python is some magnitudes > faster than Jade. > > > However, XSLT could be insufficient if the LaTeX generated is > > very complex. The loops and conditionals in XSLT are very basics, and > > variables and parameters are limited. > > I wouldn't care too much about the Python output now, HTML is important > and XSLT is THE choice here. > > > I love sablotron. It is fast and very easy to use. Also, it has almost > > every feature of the XSLT standar. > > I see, how long would it take to transform into HTML compared to Jade? The xsltproc that comes with libxslt works out of the box, too and is being used a lot in the GNOME project already. > > When you "XSLT files" you must say "XSLT file". Unlike DocBook, with > > XSLT we only will be able to produce one file, instead of one by > > <sect1>. > > > Anyway, it will be easy :) > > So we cannot slice the HTML output into several files? That's sort of a > problem, really. It took quite some time to figure out how to get Jade > to do that and still releases are a pain in the neck because there's a > lot which has do be done manually still. This is simpoly not the case. Norm Walsh has produced an XSLT package that allows chunking, as it is called. libxslt supports it and, as I understand, a some other XSLT processors do as well (it requires an extension to the engine, but the mechanism for supplying this is documented in the spec, etc). Saxon, for one, can handle the chunking also, since that is what a lot of the DocBook developers use. Basically, you can specify which level of sectioning gets put on a new page (by default, each <sect1> is a new page). You can also configure whether the files are named by using id() calls, or by using the id tag of the leading element (the latter is preferable, providing each section tag is written as <sect1 id="layers-dialog">, etc). Regards, Malcolm --
