Re: [CSS21] Issues with inline formatting model (particularly 10.8)

Peter Moulder wrote:
> I apologize for having tried to make review comments that the
> specification isn't ready for yet.
> 
> Would it be fair to say that it would be best for me not to make
> any further review comments on the specification until these
> basic issues are sorted out?
> 
> I'm happy to share my implementation experience, but it looks as if
> it's more useful for me to keep quiet until these basic issues are
> worked out.

Hi Peter,

Just to clarify, I'm not associated with the WG; I'm just an interested
party, like you.  So I speak only for me, and any claims I make about
what's in the pipeline are opinion based purely on my interpretation of
discussions held on this list or on occasional personal correspondence
with others involved.

Personally, I'd say that review comments are welcome at any time (and
I'm not sure there's time to wait until some issues are solved before
raising others, since all issues need time to review and there's
pressure to get CSS21 finished quickly).  However, in the specific case
of the boxes/elements thing which is well-known to contain deep issues,
  I suggest following fantasai's lead about the order in which to tackle
them.  For now, I believe the WG is reviewing the box-naming proposal in
[1]; the minutes will no doubt indicate the outcome.  Your comments in
that thread seemed relevant and useful, and I'm sure that further
comments would be valued.

(FWIW, I didn't believe that the boxes and elements issues would even be
contemplated for revision in CSS21, so I'm already delighted about the
progress being made.  To a certain extent, one has to set one's
expectations appropriately, bearing in mind that the spec was written in
a different era in which the demands for preciseness were much lower
than today.  There remain some aspects of the spec which I don't believe
will be fully revised; floats and tables come to mind, although I note
that tables have already been considerably improved thanks to the
dedicated hard work of several participants).)

> If this is to be my last message to www-style for a while, then
> I'll note here just for the sake of process and archived statements
> that the current text of the CSS2.1 specification is not implementable,
> that there are various issues with inline formatting model, box heights,
> table layout, counter generation, rendering etc. that mean I cannot
> reproduce other browsers' behaviour just by following the text of the
> CSS2.1 specification.  I know you're all working on fixing these and that
> doing so takes time.

Many of those issues have currently active threads and I for one think
it's appropriate to present your thoughts in those discussions now.
(It's worth reviewing the issues list[2] and scanning some of the
back-history of those conversations though, to see what's already been
raised.)  I don't recall any recent discussions on counter generation,
so it sounds like it's worth your while starting one.


[1] http://lists.w3.org/Archives/Public/www-style/2010Jul/0383.html
[2] http://wiki.csswg.org/spec/css2.1

Cheers,
Anton Prowse
http://dev.moonhenge.net

Received on Wednesday, 4 August 2010 07:29:50 UTC