Re: [CSS21] WG process - next steps for CSS21?

On Thu, Mar 03, 2011 at 03:48:49AM +0000, Linss, Peter wrote:

> The issues list we're paying attention to is at:
> http://wiki.csswg.org/spec/css2.1
> 
> If there are issues that you know about that aren't on that list,
> please let us know ASAP (and not a simple statement, like "I know of
> more issues", give us concrete pointers) so we can have the list up to
> date before our F2F.

The messages from the public sent during the working draft comments
period don't yet appear on http://wiki.csswg.org/spec/css2.1 as far as
I can see.

Note that a handful had a subject line with "[CSS 2.1]" rather than the
requested "[CSS21]"; though I didn't notice any other variations
on that sent during the working draft comment period.

In particular, we can reasonably assume that messages whose subject
contains `wd' or `working draft' (ignoring case) contain what the sender
considers to be an issue.

Is that a concrete enough pointer, or are individual message URLs needed?


Anton Prowse also mentions that

> There are *lots* of issues which were raised when the spec was CR
> that haven't been filed on the issues list

This is something I don't understand very well: the w3 pages say that

  W3C welcomes review of specifications from the moment they are first
  published until just before they are published as standards

and goes on to invite readers to "report a bug" to the place indicated by 
the "Status Section" of the document (namely www-style in the case of
CSS21).

As Anton notes, there are many such comments have duly been sent to
www-style where an issue hasn't so far been created in the main CSS2.1
issues list.

Are such messages eligible to be considered as "issues raised about the
document since the previous step" or not?  More generally, what counts as
an "issue raised about the document since the previous step" ?


> CSS 2.1 has been in the pipeline for far too long, many other specs
> normatively depend on it and it needs to be a REC to unblock progress in
> other areas. 

What other specs are those, and what bits of CSS2.1 do they depend on?
The main other spec I'm aware of is SVG, though it doesn't actually use
much from CSS beyond parsing and cascading.  It does use a few individual
properties like color and font-related properties, but doesn't use the
box model stuff.

This is significant, because most of the problems with the current CSS
2.1 text are in areas that SVG doesn't actually need.

Knowing what the dependencies are might help us concentrate on what
actually matters to other specs.


> At this point, we're going to be _extremely_ draconian about deferring
> any and all issues that can possibly be deferred. Please help us make
> this very important milestone by understanding our efforts and reasons
> for doing so. It is not our intention to ignore any issues or people's
> feedback, we'll simply move the issue to the 2.1 errata or a CSS3
> module and deal with it later. 

I agree that the problems remaining in the current CSS2.1 specification
are not easy to solve in a reasonable time frame and that it's best to
publish it as is, but given the definitions of the maturity levels,
it's important to note on the cover page of CSS2.1 that CSS2.1 cannot be
correctly and interoperably implemented just by reading and implementing
the specification document.  Working group members have been fairly
explicit about this, noting that what matters is whether the test suite
is clear enough rather than whether the text of the specification is
clear enough.

pjrm.

Received on Thursday, 3 March 2011 06:56:48 UTC