RE: [css3-multicol] pseudo-algorithm

I think getting to zero-width columns is perfectly fine. What I find odd is that the algorithm has discontinuity where you first get to zero width, then with *less* total width you get *more* width for text.

I propose that the branch that lines 20-26

	(20)    elsif (column-gap >= available-width) then
	(21)      N := 1;
	(22)      W := available-width;
	(23)    else
	(24)      N := floor(available-width/column-gap) + 1;
	(25)      W := (available-width - ((N - 1) * column-gap))/N;
	(26)    fi(20)    elsif (column-gap >= available-width) then
	(21)      N := 1;
	(22)      W := available-width;
	(23)    else
	(24)      N := floor(available-width/column-gap) + 1;
	(25)      W := (available-width - ((N - 1) * column-gap))/N;
	(26)    fi

Are replaced with a formula that keeps column width at zero but preservers column count:

	(20)    else
	(21)      N := column-count;
	(22)      W := 0;
	(23)      column-gap := (available-width/column-gap);
	(24)    fi

Or if it is not acceptable to change column gap, why not just go to 1 column???

	(20)    else
	(21)      N := 1;
	(22)      W := available-width;
	(23)    fi

-----Original Message-----
From: Håkon Wium Lie [mailto:howcome@opera.com] 
Sent: Tuesday, February 08, 2011 11:10 AM


But implemenations have to deal with zero-width columns anyway; if you set:

   column-count: 999999999

or something, you will effectively have zero-width columns. 

Given that we clip columns, I think it's managable. 

Another alternative is to try find some minimal column width, as suggested here:

  http://lists.w3.org/Archives/Public/www-style/2010Nov/0434.html

However, I don't see any simple ways to improve such overconstrained situations. Feel free to propose something, though.

-h&kon
              Håkon Wium Lie                          CTO °þe®ª
howcome@opera.com                  http://people.opera.com/howcome