W3C home > Mailing lists > Public > www-style@w3.org > November 1997

Re: Hey Microsoft! cool it with CSS points ok?

From: Todd Fahrner <fahrner@pobox.com>
Date: Fri, 24 Oct 1997 15:24:32 -0800
Message-Id: <v03102807b076ddc7392e@[206.245.203.103]>
To: "Eric A. Meyer" <eam3@po.cwru.edu>, www-style@w3.org
Eric A. Meyer wrote (5:54 PM -0500 11/24/97):

"    Okay, then, are my suggested properties (see previous post) all that
" unreasonable?  Especially if UAs can define the meaning of
" 'font-size-minimum: legible;' to be a given pixel height.  On a Mac, it
" could be 9px; on Windows, it could be... whatever.  I mean, if this is a
" terrible idea, rip it apart.  Otherwise, should we actually be considering
" it?

I think it's a reasonable address to a problem that could be solved in a more general fashion, without having to add new features to the unimplemented spec. 

Assume a perfect CSS1 implementation - user stylesheets and everything. Clueful CSS authors specify all sizes for all elements (objects/images too) in em units, descending from a single value on BODY. Authors understand that specifying sizes much below 1em (whose real value is inherited from BODY) is risky business. Why? Because users define the value of 1em on BODY in their personal style sheets as "comfortable" and attach !important weight to it. Thus all type sizes in the document are anchored to the user's "comfortable" level. If authors screw up (as they will), and spec type well below 1em, (say 0.5em), users can recover by overriding the value on body upward, perhaps by a slider or similar UI element. This just writes a new value to the user stylesheet. 

Note that the full range of _relative_ sizes specified by the author is preserved in this scheme, which may simply be necessary to differentiate the 12 levels of heading in, say, HTML 5.

"    No, I don't; at least, not about the graphics.  I was thinking of text
" headlines getting too big to fit into whatever region of the screen they
" had available, even with soft wrapping of the text... which is one of the
" reasons (other than simple symmetry) I suggested a 'font-size-maximum' to
" go with '-minimum.'

I think it's bad to clip the allowable range. If headline type is so large that even hyphenated syllables of individual words don't fit in the window or column width.... I think we're talking about a pretty extreme case aren't we? Not one I see as a burning problem really. Maybe the piece in question is a type specimen, where 500-pt type is appropriate?

" >The disconnect occurs because not all
" >elements' sizes are specified in the same unit system.
" 
"    I understand your point.  However, in my visions of an ideal Web page,
" graphics aren't a big part of the interface, they're pleasing (but
" basically content-empty) decorations.  

O I'm with you about not liking fluff graphics - decoration. I think typography itself is the loveliest of the graphic arts, though, so don't draw too hard a line. 

"    I agree with you, as it happens.  I try to make my documents as portable
" as possible, even across media types.  But if we're going to encourage
" authors to use relative measurements, knowing that this might push text
" into illegibility, should we consider offering them the ability to set
" floors and ceilings?

I'd say let the users raise the floor when necessary, but let the ceiling take care of itself. Most often, it will.


Todd Fahrner
mailto:fahrner@pobox.com
http://www.verso.com/agitprop/

The printed page transcends space and time. The printed page, the infinitude of books, must be transcended. THE ELECTRO-LIBRARY.
	- El Lissitzky, 1923
Received on Monday, 24 November 1997 18:19:59 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:53:52 GMT