Re: New layout language.

> I guess you mean the code defined in
> http://w3.org/mid/abd6c80105070617165f3f9813@mail.gmail.com
> 
> The fact that CanvasWidth and CanvasHeight are defined outside
> Form1() constructor makes it harder to decipher the fact out of the
> code. Adding a couple of comments wouldn't be too bad.

Yeah, I was in a hurry. Sorry.

> > The canvas represents the initial values for the canvas to be compared
> > against the current size of the viewport. The code is pretty simple.
> > Take a look at it.
> 
> I'll try to explain this in my own words and we'll see if you agree
> with my interpretation.
> 
> All regions are computed as [left|top|width|height]Growth was zero.
> Then a bounding box containing all regions is computed to get canvas
> size (I'd prefer this method to compute the canvas size even though
> you used hardcoded canvas size in your example implementation - this
> guarantees no extra whitespace around the layout). Then the
> *remaining space* (sounds a bit like %%-unit, doesn't it?) is shared
> between different regions according to [left|top|width|height]Growth.

Actually canvas size is specified which gives you your right and
bottom margins. However it could also be done with a single
right-margin/bottom-margin pair that would look to the rightmost and
bottommost regions. Including this simple value get around the need
for right and bottom. Otherwise it's pretty much correct.

Either bottom/right padding would need to be implemented or
width/height needs to be implemented, but only for the canvas. I
prefer width/height because it reduces the number of concepts needed.

Yeah, I forgot to thank Andrew for the basic idea. He has been so
persistent that it's been in my mind for awhile now. He had
implemented it in his browser, but people kept saying it couldn't be
done. So I guess percentage free space was keeping me occupied.

-- 

Orion Adrian

Received on Thursday, 7 July 2005 14:02:23 UTC