W3C home > Mailing lists > Public > www-html@w3.org > June 2005

Re: separator abuse

From: Orion Adrian <orion.adrian@gmail.com>
Date: Wed, 1 Jun 2005 08:03:06 -0400
Message-ID: <abd6c80105060105036b706cb2@mail.gmail.com>
To: www-html@w3.org

Structurally I would say that most documents, fiction and non-fiction
do fall into a neat tree structure. However once you leave the outline
that produced the table of contents it doesn't anymore. Separators are
content, not structure. I really do like the "Meanwhile at..." analogy
that was presented.

Character styles and table styles do not fit nicely into a tree structure. 
Not every element has a logical parent.
Not every element has a logical child.
Some content is best represented by empty elements, replaced at render time.

It's almost like the model for paragraph shouldn't be XML, but rather
another grammar. In this case a non-tree grammar.

Tables however can be hacked rather effectively into a tree structure;
they just aren't. Nicer CSS selectors would be beneficial here.
However CSS doesn't allow for the extension system that it should.
Each document type should be allowed to specify it's own
pseudo-classes and pseudo-elements and specialty properties.

Orion Adrian

On 6/1/05, Mikko Rantalainen <mikko.rantalainen@peda.net> wrote:
> Edward Lass wrote:
> > You mean verses might not fit neatly into paragraphs? You mean the Bible
> > doesn't consist of hierarchical information?
> >
> > You mean one of the most important historical works in human history
> > (to say nothing of its spiritual significance for many) can't become
> > well-formed XML?
> It can. But if we accept the fact that the Bible doesn't have any
> logical connection between paragraphs and sections we cannot markup
> it as a tree where a paragraph is always inside a section.
> And as Johannes Koch already mentioned, this multi-hierarchical
> structure is no different from musical notation with notes, bars,
> beams and slurs.
> Either you come up with a new markup languages MusicML and BibleML
> to somehow represent this kind of structure or you cannot markup the
> whole thing with semantic markup.
> One could represent the Bible as a sequence of <section> elements
> that contain only <l> elements. A paragraph change in the Bible is
> represented by an empty <l> element. It's not a nice semantic coding
> of the Bible but it's still well-formed XML.
> There must be some rules that we all have to adhere to or there's no
> way to markup a document with a semantic markup. XHTML2 already
> defines things such as headers, sections and paragraphs. If you come
> up with some historical document that has no logical connection
> between these things, there's no point trying to mark up that
> content with the elements defined in XHTML2.
> A piece of text inside XHTML2 file should be surrounded by <section>
> and </section> tags only and only if it matches semantically the
> <section> element's semantics defined in XHTML2. The semantics of
> <p> element defined in XHTML2 (or any version of HTML) doesn't match
> the semantics of paragraph in the Bible so it's logically an error
> to try to use that markup for a paragraph of the Bible.
> --
> Mikko
Received on Wednesday, 1 June 2005 12:03:09 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:06:10 UTC