You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@forrest.apache.org by "Piroumian, Konstantin" <KP...@flagship.ru> on 2002/03/13 19:07:28 UTC

[DRAFT] home.xml

Good evening! (Yes it's already evening here)

Just want to check if I move in the right direction. In the attached file
you'll find main content part of home.html in XML representation.

During writing that file I had a feeling that XML/XSLT is not a very good
choice for this kind of pages. Look, a rather simple HTML page requires more
than 3 DTD/Schema (menu, contents, sections), one or more XSLT stylesheets,
Cocoon - to connect all it together.
Does it worth so much instead of having one or two versions of HTML? What if
the number of pages with unique layout will be high?
I'm not serios, of course, but many a true word is spoken in jest (its
English equivalent for Russian proverb).

I'm waiting for feedback to start the XSLT part.

Regards,
    Konstantin


RE: [DRAFT] home.xml

Posted by Steven Noels <st...@outerthought.org>.
Konstantin wrote:

> Good evening! (Yes it's already evening here)
>
> Just want to check if I move in the right direction. In the
> attached file
> you'll find main content part of home.html in XML representation.
>

Thanks!

I don't have the time now to go into much analysis, but I'm wondering
why we should invent elements for the content-part of the page: nothing
is there which we shouldn't be able to mark up using plain xdoc
elements...?

If we really want to have a different representation for the same xdoc
content elements depending on the section/page in which they appear, we
should vary the stylesheet and not the markup. Which means the content
part of the aggregated pages should be based on the body element of
document.dtd.

Now that of course means we should try to re-use as much elements
(especially body elements) across the different doctypes (document, faq,
changes, specification, javadoc and todo), and ideally we should do so
using externally declared parameter entities ( yes - those pesky
entities again, even in the light of XML v2 :-( ). A brief look into the
existing DTD's in the repository tells me that they are pretty
monolithic, i.e. they do not have much common elements (which is good,
perhaps we don't need common elements after all, although attributes is
another case).

Thinking along while writing: I believe the content part of the
aggregated page should consist of the body-type elements of each
'authoring' DTD. That being said, the schema for such a thing will be
quite hairy if we don't put all this body-type elements in a external
entity to which we can refer from all authoring DTD's and the schema for
aggregated pages.

Another solution might be to have a set of stylesheets which transform
the content docs into something like XHTML prior to aggregation.

What should we do?

</Steven>