Re: Seriffed and sans-serif fonts?

Brian Behlendorf (
Sun, 23 Apr 1995 23:41:31 +0500

On Sat, 22 Apr 1995 wrote:
> At the WWW conference in Darmstadt, there was a short session in the
> Developer's Day about style sheets.
> One thing that came out is that there may be situations where the author must
> be able to override the user's style sheet preferences -- e.g. for legal
> liability reasons.

Interesting - I looked at the Darmstadt pages and couldn't find notes on
this - could anyone elaborate?

> Another is that people are actively working on DSSSL Lite, which is intended
> to be a subset of the complex ISO DSSSL standard that's easier to implement
> efficiently enough for a browser.
> It's also clear that once you have a browser that doesn't have a fixed
> mapping from tags to styles, you might as well make the tag list extensible;
> it's a short step from there to supporting DTD-driven SGML documents.
> At least two SGML browsers are coming shortly, one from us (SoftQuad Panorama)
> and I _think_ there's at least one other SGML WWW viewer/browser in
> development. Several people mentioned to me that they were thinking of it,
> although I'm thinking of becoming a multi-millionaire, so that shouldn't be
> taken too seriously yet.

This is all definitely as things should go, but I worry that style sheets
won't even be introduced as a possibility until SGML-based browsers are
out there and used in abundance, which could be a long time. If we can
decide on a general mechanism for how they are referred to (using a
<STYLE> tag to embed it directly, or <LINK> tags in the HEAD to pull them
down when needed, or...) then browser authors can start trying them out
sooner. Who's working on the DSSSL Lite proposal? The "experimental
style sheets" link off of Arena's home page is giving me a 403 error....

> There are some problems with the current style sheet syntax when included
> in a document -- HTML 2 browsers either go bananas or display the text
> of the style sheet at the top of the document -- but I think that could be
> fixed by putting the syntax into elements with attributes.

But, HTML 2 browsers should never have to see them, if the <STYLE> tag is
HTML 3 only :) If one uses a <LINK> mechanism, then it's not a problem.

Hmm - seems that if we use <LINK>, stylesheets should have their own mime
types, be it text/stylesheet, or even more elaborate like
text/dsssl-lite-stylesheet, or maybe stylesheet/dsssl-lite? That way we
could have different stylesheet syntaxes with their own specifications,
and as long as they were defined in terms of SGML then they might be able
to coexist peacefully...


--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-- http://www.[hyperreal,organic].com/