- From: Nir Dagan <nir@nirdagan.com>
- Date: Wed, 10 Nov 1999 13:36:09 -0500
- To: Piers Williams <PiersW@zinc.co.uk>, "'www-html@w3.org'" <www-html@w3.org>
this is already taken care of in Cascading Style Sheets level 2: http://www.w3.org/TR/REC-CSS2/visudet.html#min-max-widths At 05:49 PM 11/10/99 +0000, Piers Williams wrote: >At the moment the HTML spec provides for developers to specify either a >fixed, or variable width for certain block elements (Tables, DIV's etc...). >However this model has serious limitations, which are a cause of constant >irritation to myself and the designers I work with. > >If a table, for example, is set width="100%", then it is up to the >user-agent to scale the table according to the agent's window. This is good, >because it enables the content to maximise use of the available space on the >user's monitor, whatever the screen size and resolution. However - given a >particularly wide monitor - lines of text can be drawn out to a width that >makes them less legible (not to mention ugly). > >Given this problem, many web designers opt for the fixed-width table model, >where an absolute pixel width is supplied for the table. This prevents lines >of text becoming too long (horizontally), but in a highly inflexible way. To >avoid the trauma of horizontal scrollbars, it is usually necessary to fix >the width of the table at 620px, which means on a large monitor over half of >the screen can remain unused. > >What is needed is a way of specifying the width of these elements in such a >way that the user-agent is free to resize them up to a point - i.e. provide >a mechanism for resizing with constraints. > >My suggestion would be to implement the attributes MINWIDTH and MAXWIDTH for >the TABLE tag (and indeed any other tag with a WIDTH attribute). The >attributes would take the same values as width (length, percentage, auto, >inherit), and would specify an lower and upper bound to the user-agent's >freedom to resize that content. > >The user-agent would only apply these attributes if both were specified, and >if so ignore the corresponding WIDTH tag. Backwards compatibility could be >preserved by specifying a WIDTH tag as a fall-back to earlier browsers. > >I would suggest that the MINWIDTH and MAXWIDTH attributes wouldn't have to >have the same value type (pixel vs. percentage etc...), but that would be up >to the user-agent to ensure that the MINWIDTH and MAXWIDTH values didn't >overlap. What would happen if overlapping MINWIDTH / MAXWIDTH values were >supplied (either directly, or indirectly via resolution of percentage-widths >into pixel-widths) is - I think - a topic for further discussion. My present >thoughts would be that the user-agent would reject invalid MINWIDTH / >MAXWIDTH pairs, and fall back on the WIDTH tag (if present). > >I would also suggest that MINWIDTH and MAXWIDTH attributes be added to the >CSS spec in the appropriate places. > > >(PS: Apologies if this has already been suggested - I can't find it >anywhere) > >Piers >-- >Z I N C >http://www.zinc.co.uk <http://www.zinc.co.uk/> > > Regards, =================================== Nir Dagan Assistant Professor of Economics Brown University Providence, RI USA http://www.nirdagan.com mailto:nir@nirdagan.com tel:+1-401-863-2145
Received on Wednesday, 10 November 1999 13:34:14 UTC