Re: The style agenda


 > | I imagine this document going through several phases. Starting as a
 > | scratchpad, it should end up as an I-D. I've changed the wording on
 > | the top to better reflect this, thanks!
 > If it's going to be an I-D, does that mean it has to come under
 > the aegis of some WG?

I would think so. Do you have a favorite?

 > It was W3's role in standardization that was confusing.  As I understand
 > your response, W3 is taking no such role; that's okay.

At this point in time, in this area, W3C hasn't taken on any
standardiszation role.

 > | Agree, in principle. Conflicts arise when you want to specify style
 > | for non-HTML information, e.g.
 > |  - the whole document: "doc: margin.left = 50pt" -- I'm not sure
 > |    we can convince people to use "html of "body" instead of "doc".
 > |  - browser buttons: "browser: button.save = off"
 > |  - html source: "source: font.size = 14pt"
 > | These conflicts will arrive when we try to handle a DTD with <SOURCE>.
 > | So, an item for discussion should be to what extent, if any, the style
 > | sheet should be able to influence non-HTML properties, or take
 > | non-HTML information (like AGE and LANGUAGE) into account when
 > | rendering. The author of a document shouldn't decide how e.g. the html
 > | source should be displayed, but certainly the user should. And if it's
 > | not in the style sheet, where will in be? I hate .Xdefaults!
 > Margins aside, I think such things go in what I think of as another style 
 > sheet, though this could be info in HEAD or a part of the rendering style
 > sheet.  There's a large iceberg of nonrendering info people want to
 > convey to the browser (browser buttons being a good example).  

Except for the button example, the others are rendering issues. As for
buttons, choosing to display or hide them is also a rendering issue
since you gain/loose screen real-estate depending on your choice. I
would hope the style sheet mechanism would convey rendering
information peripheral to HTML as well. Users will ask for it.

Are you suggesting we should have two style sheet mechanisms?

 > I don't know about the current mechanism, but let me put the case for
 > something like "legal."  As a publisher, I may need to insist that 
 > my doc be read only according to a style sheet I supply.  I know that
 > I probably can't find a mechanism to enforce that on the browser,
 > but at least I have to be able to say it.  (Then if the user insists
 > on rendering Warnings as gray text on a grey background, fails to
 > read a Warning, and electrocutes himself, I'm in the clear.)

This seems to be the prevailing view among publishers, and it makes
some sense. But, I don't think the style sheet mechanism is the best
legal cover you can find. Since style sheets can be LINKed from a
document (most proposals suggest this), it may or may not reach the
browser. So, should the browser refuse to show a document just because
the style sheet may include lagal hints? I don't think so. This is
just one of the questions that arise if we decide to deal with

 > | I'm proposing a simple stream of boxes where the style sheet controls
 > | the margins of each box. The boxes are stacked on top of each other
 > | (perhaps with a page break here & there), and children inherit the
 > | parent's horizontal boundry + margins (i.e. to do indented nested
 > | list). Paper delivery will require some extensions to this model.
 > Could you compare this model to Tex's?  No glue, obviously ...

No glue. And no nested boxes. Just a simple stack of boxes similar to
how current browsers format HTML (tables and images in text flow

|               |
|*P             |
|               |
|@@UL           |
   |            |
   |%LI         |
   |            |
   |%LI         |
   |            |
   |@@UL        |
      |         |
      | %LI     |
|               |
|*P             |


 '*'  = p  [margin.left]
 '@@' = ul [margin.left]
 '%'  = li [margin.left]

Children inherit their parent's x position + margin.left.

Thus, one can control list indentations by setting margins. 

Text flow around floating images does not fit well into this model,
e.g. you probably don't want to use the normal margins when placing
text next to an image.

One could add some extra properties if control at that level is


Hakon W Lie, WWW project CERN, CH-1211 Geneva 23