Lists Home |
Date Index |
David Carlisle writes:
> don't you have cross reference elements where you need to generate
> text by chasing the reference?
After years of doing it, chasing the reference myself is pretty easy,
and I rarely generate text that way.
> or require tables of contents
When books go into production, yes, that's necessary, but it's hardly
necessary when reading chapter by chapter. I tend to start the
editorial process with a proposed TOC, but generating TOCs to see where
we are is rarely necessary.
> and other
> generated lists and navigation links
I guess I tend to generate such things once and leave them in the
document. I don't have a lot of going back and forth between original
source and final document. Sticking with a relatively simple set of
conventions gets me to the 80% point easily.
> or have " highly structured
> XML representing data fields that you want to display in some
> different order in natural language paragraph style?
That's the 3% of cases where I need something else for presentation.
> I always seem to need to do all of these things (before we even get
> to worry about the mathematics)
Some people seem to need those things, some people don't.
I'm not entirely sure why the early "CSS is for HTML, XSL is for XML"
story was taken so seriously, except perhaps that HTML people had long
been used to the constraints of document order while more XML people
wanted the advanced capabilities.
To me, transformation's neat, but it's far past my 80/20 point for Web
presentation and even casual print presentation.
> CSS is fine if you just want to syntax highlight your XML but you
> really want to see the structure of the XML file (perhaps because you
> are editing it) rather than see the document resulting from that XML.
CSS display properties do perfectly well for my needs in that regard.
There support for blocks, inlines, lists, tables - it's quite useful for
> that's just saying that on the web most people are familiar with html,
> this is of course true but it doesn't have to be that way. Obviously
> there will always be (perhaps a majority) of such documents, but xhtml
> is fine for them, xml as "sgml-on-the-web" was supposed to promise
> the possibility of putting other kinds of documents up on the web as
Even beyond the Web I think far more people are familiar with documents
whose presentation order is the same as their typing order. HTML has
it, but so does WYSIWYG, typewriters, etc.
I'm afraid that I see people who genuinely need transformations for
presentation and are willing to learn about transformation into order to
create such presentation are in a distinct minority. A very helpful
minority, to be sure, useful in hard cases, but certainly not ordinary.
> > Their handling of the XHTML namespace and mixtures of XML and XHTML
> > is execrable to put it mildly;
> oh yes teh html side and its xml islands is _not_ what I was talking
> about, or use. But sending XML as tex/xml to the XML side of IE,
> styled via XSLT works quite well (even if the document's full of
And why should anyone HAVE to go to that trouble, if they didn't
actually NEED a transformation?
> > And the CSS in Mozilla and Netscape (and to a lesser extent in
> > Opera) puts IE completely to shame for any XML application.
> But I only ever use css for some simple colouring or font choices,
> that works well enough and unfortunately opera doesn't have any
> transformation possibilities last time I checked so isn't usable on
> any of the xml I see.
And IE has crap CSS support for XML, so I don't consider it usable on
any of the XML I see.
> > I'm happy with CSS for pretty much all the direct DocBook reading I
> > do up to the point where ORA feeds it all into Frame.
> as I say above css styling xml is fine if you just want a draft
> editing view, but could you really get a final presentation form for
> use by someone other than the author without using a transformation
Yes you can, depending on your needs.
Simon St.Laurent - SSL is my TLA
http://simonstl.com may be my URI
http://monasticxml.org may be my ascetic URI
urn:oid:18.104.22.168.4.1.6320 is another possibility altogether