RE: [CSS21] Ambiguity in tokenizer, "normative appendix G"

On Thursday, February 5, 2015 1:56 AM Florian Rivoal [mailto:florian@rivoal.net] wrote:
> 
> > On 05 Feb 2015, at 00:00, Tab Atkins Jr. <jackalmage@gmail.com> wrote:
> >
> > On Thu, Feb 5, 2015 at 7:03 AM, Florian Rivoal <florian@rivoal.net> wrote:
> >>> On 04 Feb 2015, at 20:50, Tab Atkins Jr. <jackalmage@gmail.com> wrote:
> >>> I don't think this is or should be our goal at all.  The number of
> >>> Level 1 specs suggests that CSS is far wider than the CSS2 base
> >>> we're still building off of.
> >>
> >> Calling it CSS 3 sounds wrong, but if we start including links to the
> (sufficiently mature) level 1 modules, it's a good document tying together all
> the mature parts of CSS, and giving an intro to the whole.
> >>
> >> At the same time, I am not what the difference is between that and the
> snapshot (assuming we maintain it).
> >

We need to do a better job maintaining the snapshot. We once committed to doing a yearly snapshot after TPAC each year but no one has done that. It needs to be on the TPAC agenda for each year so we can all agree what is really sable enough for authors. If it doesn't change we still publish the note just with a new date on it.

> > Yeah, you've just described the snapshot.
> 
> The thing is, if we maintain CSS2.x and the snapshot, it seems to be that
> they'll be converging in content and purpose, so I am wondering if they
> should (eventually) merge.
> 

The snapshot will not converge with 2.x until level 3 specs have completely replaced all 2.x sections. The snapshot contains level 1 specs that are not part of 2.1 and will never be part of 2.x. Therefore there might never be a merging (at least not the prose from 2.x). The prose from 2.x should be replaced by level 3 spec prose that replace the particular sections of 2.x.

I am also not really against a merging in the future if we see more of a need for it. I just am not seeing it right now or in the near future.

> For now, CSS2.x has normative prose, and a nice intro to the language ()
> which and the snapshot doesn't. The snapshot has links to mature specs
> other than 2.1, which 2.1 doesn't do.
> 

The normative prose are specifically for 2.x. The normative prose for all the other specs should be within their respective documents as well. The snapshot just says what is stable enough for authors to use.

> But as we maintain both, the number of CSS2.x sections superseded by
> something will grow. Then, one of the following happens to CSS2.many:
> a - It's full of irrelevant text (because we left it there) b - It shrinks until
> there's only the intro there (because we deleted all the superseded stuff) c -
> It shrinks until there's only an intro, section headers and pointers to newer
> specs.
> 

This assessment seems right...

> I think 'a' is bad.
> If we get to 'b', we might as well inline the intro into the snapshot.
> If we get to 'c', CSS2.many and the snapshot will be mostly the same thing,
> except the snapshot will also have links to modules that are additions (not
> just replacements), while CSS2.many will have a nicer intro. At which point it
> feels like they should merge.
> 

'a' is bad for sure.
'b' isn't quite right either. The snapshot is a note containing specs that are finished (REC) or are stable enough for authors to really use without too many more changes or interoperability problems.
'c' I think if we get to this stage, which I think will take years, we might want to consider pointing more people at the snapshot and maybe 2.many isn't even necessary at that point. There probably would be no need to merge.

> I am not saying we need to merge them both *now*, but it seems that the
> long term trend is for this two things to get increasingly similar, and we
> should probably consider their evolution together.
> 

I don't think we are at a point where it makes sense to merge these now I agree.

> Or am I wrong, and to we expect that some normative parts of 2.1 will never
> be replaced, and that we won't get into the scenario where 2.x shrinks to just
> the introduction (+links)?
> 

In the end 2.1 should be completely replaced if it isn't then we have missed something in our newer specs.

--
Thanks,
Arron Eicholz

Received on Thursday, 5 February 2015 15:21:56 UTC