RE: several messages

At 2001-06-29T16:07+0300, Manos Batsis wrote:-

> > -----Original Message-----
> > From: Ian Hickson [mailto:ian@hixie.ch]
>
> > This has never been a problem with HTML as far as I am aware;
> > why would
> > it suddenly become a problem with XML?
>
> Sorry Ian, but I cannot share your point on this one.
>
> In HTML, the agent knows what goes where (since it's a presentation
> centric language) while the structure of an XML document doesn't help a
> browser in deciding the presentation structure that will make sense to a
> user. So, either the XML document should have a structure and 'data
> order' close to the desired XSL output, or an XSL is mandatory while CSS
> on it's own is useless. So, an agent should know what is what IMHO.

I'm not sure I follow your argument. Are you saying that the type should
be required in the markup because one type of style sheet might be more
important than another? How would the UA know? I don't think anyone is
arguing that having the type information in advance would not be useful,
but it should rarely be a great inconvenience to determine it by other
means.

But I think Ian's comment was about authoring tools, where it is probably
even less of a problem. Broadly, I see two possibilities: 1. The authoring
tool has at least some knowledge of style sheets, in which case it should
already be able, for example, to add a style sheet to a document without
the author having to tell it the media type explicitly. 2. It knows
nothing about style sheets, in which case the presence or absence of type
information is irrelevant to it.

HTTP content negotiation can be extremely useful. There would have to be a
very good reason to prevent its use.


Tim Bagot

Received on Friday, 29 June 2001 10:18:00 UTC