W3C home > Mailing lists > Public > www-style@w3.org > February 2011

RE: [css3-multicol] pseudo-algorithm

From: Stephen Zilles <szilles@adobe.com>
Date: Tue, 8 Feb 2011 11:23:45 -0800
To: Sylvain Galineau <sylvaing@microsoft.com>, Alex Mogilevsky <alexmog@microsoft.com>, Håkon Wium Lie <howcome@opera.com>
CC: "www-style@w3.org" <www-style@w3.org>
Message-ID: <CE2F61DA5FA23945A4EA99A212B1579537CDA58233@nambx03.corp.adobe.com>
I guess I do not remember discussing the fallback rules when multiple columns do not fit, but just thinking as someone who wants his content to show-up, I would object to either having zero width columns or zero-width column-gaps. 

It seems to me that when column gaps are causing the problem (i.e., are leading to zero width columns), then the fallback should be to a single column. (I think I am agreeing with Alex, here.) Unfortunate, the example we have been discussing seems to be sufficiently unrealistic for me to understand why it is important to preserve the size of the column gap. Is there a use case that would illustrate this importance?

Also, is there a use case that says it is important to preserve the number of columns? If not, then falling back to a single column seems like a good solution. Better even than having two single character columns and a preserved column-gap.

It would seem to me that what is needed is a min-column-width and a min-column-gap to insure that both can shrink and that they will never shrink to something silly. This, would complexify the pseudo-algorithm, but would give more IMHO reasonable results.

Steve Zilles


> -----Original Message-----
> From: www-style-request@w3.org [mailto:www-style-request@w3.org] On Behalf
> Of Sylvain Galineau
> Sent: Tuesday, February 08, 2011 11:02 AM
> To: Alex Mogilevsky; Håkon Wium Lie
> Cc: www-style@w3.org
> Subject: RE: [css3-multicol] pseudo-algorithm
> 
> I was wondering about that. Taking another look
> at this branch of the algorithm it certainly seems
> to be designed to preserve the specified column-gap,
> no matter how large it may be in relation to the
> available width. Is that intentional ? Adjusting
> the column-gap may complicate things somewhat but
> do author really give the gap size a higher priority
> than column-width/column-count ?
> 
> > -----Original Message-----
> > From: Alex Mogilevsky
> > Sent: Tuesday, February 08, 2011 10:54 AM
> > To: Håkon Wium Lie
> > Cc: Sylvain Galineau; www-style@w3.org
> > Subject: RE: [css3-multicol] pseudo-algorithm
> >
> > Ah, I see, after you fix line 24, changing line 17 to "<=" makes no
> > difference when zero-width columns with their gaps fit precisely into the
> > width.
> >
> > I may have missed the discussion that led to lines 23-26 though. That
> part
> > of the algorithm takes off at the point where nothing fits at all and
> > tries to give priority to column gap over column count, which leads to a
> > transition from 3 zero-width columns to 2 usable columns. 2 columns could
> > be useful, but if you have to hit zero first I find that questionable...
> >
> > It seems to me there shouldn't be any way to get from 3 columns to 2
> > columns. The algorithm should either decrease column gap to keep 3 zero-
> > width columns indefinitely, drop all the way to 1 column or produce
> > overflow.
> >
> > -----Original Message-----
> > From: Håkon Wium Lie [mailto:howcome@opera.com]
> > Sent: Tuesday, February 08, 2011 6:25 AM
> > To: Alex Mogilevsky
> > Cc: Sylvain Galineau; www-style@w3.org
> > Subject: RE: [css3-multicol] pseudo-algorithm
> >
> > Also sprach Alex Mogilevsky:
> >
> >  > BTW you can further improve this degradation track by changing line
> > (17) from  >
> >  > 	(17)    if ((column-count - 1) * column-gap < available-width) then
> >  > to
> >  > 	(17)    if ((column-count - 1) * column-gap <= available-width) then
> >  >
> >  > If you consider zero-width columns an acceptable multicolumn  >
> layout,
> > you will get 3 zero-width columns at 50px gap.
> >
> > To me, it seems that changing line (17) will have no impact; you will in
> > any case get 3 zero-width columns at 50px gap. The only differnce is that
> > line 18 will set it to three, rather than line 24.
> >
> > I think we must allow zero-width columns. This will also occur when
> > 'column-count' approaches infinity. But we probably don't want to allow
> > 'column-width: 0'. In the editor's draft, I have clarified that zero is
> an
> > acceptable computed value for 'column-count'.
> >
> >    http://dev.w3.org/csswg/css3-multicol/
> >
> > -h&kon
> >               Håkon Wium Lie                          CTO °þe®ª
> > howcome@opera.com                  http://people.opera.com/howcome
> 
Received on Tuesday, 8 February 2011 19:24:38 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:37 GMT