W3C home > Mailing lists > Public > www-style@w3.org > December 2005

Re: styling xml with css - copying xml attribute values into CSS attribute values

From: Orion Adrian <orion.adrian@gmail.com>
Date: Wed, 14 Dec 2005 16:06:04 -0500
Message-ID: <abd6c8010512141306g373c57ddq404319f4a680095@mail.gmail.com>
To: www-style@w3.org

On 12/14/05, Noah Scales <noahjscales@yahoo.com> wrote:
> Hi, Orion.
>
>  You wrote: "What XHTML does do is describe the relationships between
>  the paragraphs, images, headings, sections, and the document/page
>  itself."
>
>  That's what I meant by "display semantics". XHTML tags are interpreted by
> browsers as having specific CSS display properties implicitly. Custom
> mark-up requires explicity CSS display properties.
>
>  You wrote: "XHTML doesn't represent the meaning of the data, just the
> structure used to communicate that data.
>
>  So what is it you're really asking for?"
>
>  Let people write their own hypertext languages. Let them use CSS to specify
> what those languages mean to browsers, to search engines, and to people
> interested in the display semantics of that mark-up.

I'm with you for browsers since they're simply presenting the data.
I'm not sure what browsers care though. A browser doesn't particularly
care about what the document means, merely what words are present and
how it links to other resources. Why would Google or any other search
engine care that <mydogsname>Bob</mydogsname> is displayed as a block?

>  Custom hypertext languages will:
>  - meaningfully describe the content they contain.
>  - express relationships closer to those in docbook, opendocument, and the
> ms word document xml.
>  - provide hypertext information by CSS mark-up to machines and spiders.

Custom hypertext languages will:
- meaningfully describe a small number of relationships.
- express relationships in meaning, rather than relationships in
structure (I like this)
- will not provide hypertext markup information

Specifically on the first and third points, I owe you an explanation:

Each format describes a small number of relationships and objects.
Also they can only describe what you tell them they can describe. So I
might say <travelrequired>yes</travelrequried> and that says
essentially the same thing as "Travel is required.", but the sheer
number of objects and relationships involved prevents any markup
language from describing more than a relative few objects and
relationships.

One the third point, CSS doesn't specify hyperlinks. I see the
problem, but it's a problem in URL design, not in CSS. URLs make
specifying queries painful. What you're looking for is the ability to
say, "I want details on this object", specifying only the object in
HTML, and not the base URL part that gets you there. Well it's also a
flaw in HTML usage/design in that it combines content and UI.

It's also a shame to me that CSS doesn't use a more
appropriate/cleaner layout algorithm, which could also help mitigate
some of the problems.

>  They'll also be easier to read and to write than XHTML.


--

Orion Adrian
Received on Wednesday, 14 December 2005 21:06:12 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:42 GMT