W3C home > Mailing lists > Public > www-style@w3.org > October 2007

RE: [becss] "Behavioral Extensions to CSS" no longer an appropriate name

From: Ian Hickson <ian@hixie.ch>
Date: Tue, 23 Oct 2007 01:11:28 +0000 (UTC)
To: Chris Wilson <Chris.Wilson@microsoft.com>
Cc: "Paul Nelson (ATC)" <paulnel@winse.microsoft.com>, "www-style@w3.org" <www-style@w3.org>
Message-ID: <Pine.LNX.4.62.0710230025080.16360@hixie.dreamhostps.com>

On Mon, 22 Oct 2007, Chris Wilson wrote:
>>>
>>> If this spec is called "Behavioral Extensions to CSS," it begs the 
>>> questions of "what's the relationship of this and Microsoft's 
>>> behaviors, heavily used and supported in 80% of the web browsers 
>>> currently in use?
>>
>> It supplants them.
> 
> Um, as in "to replace (one thing) by something else" [1]?  No it 
> doesn't, because it doesn't work in IE <= 7 at the very least, and 
> 'behavior' does; and I'm unaware of any commitment by Microsoft to 
> remove 'behavior' and implement 'binding' for the same purposes.  In 
> spirit, I can see how this is intended to supplant 'behavior' - but 
> please don't reuse the same specification for something that is beyond 
> substantial different (and borders on totally different).  That seems 
> revisionist to me.

The new BECSS and XBL drafts really are the latest step along the same 
line of development. No revisionism is intended.

XBL and HTC have a _lot_ in common. Indeed, XBL1 has been used to 
implement HTCs in Mozilla [1]. The primary differences are that XBL 
doesn't require a whole new browsing context per behaviour instance, that 
the processing model is defined well enough that we can (in theory at 
least) get independent implementations, and that the format uses XML 
syntax and integrates with the DOM and CSS.

[1] http://dean.edwards.name/moz-behaviors


> >> Why isn't this just the 'behavior' property that Microsoft 
> >> implemented nearly a decade ago, since it seems to do the same 
> >> thing?"
> >
> >Because it is instead the 'binding' property that Mozilla implemented 
> >nearly a decade ago, which also seems to do the same thing, but happens 
> >to have a few advantages (e.g. its vendor wrote a more detailed spec).
> 
> Mmm-hmm.  Had anyone expressed interest in implementing, or indeed even 
> given feedback about BECSS, we would have written a more detailed spec 
> as well.

Since, as you imply, nobody expressed interest in implementing, or indeed 
even gave feedback about the original BECSS, we can presume that it is in 
fact mostly irrelevant at this point. It's unclear to me why you are 
worried about it.

Note that there is still a permanent URI for Microsoft's HTC submission:

   http://www.w3.org/TR/1998/NOTE-HTMLComponents-19981023

The XBL2 spec acknowledges it explicitly. The only thing that is really 
"lost" in reusing BECSS is Action Sheets.


> As it is, the original BECSS is roughly twice as long as this 
> new draft?

Length doesn't necessarily correlate to detail, but in this case the 
shortness is mostly due to the fact that the bulk of the spec is actually 
in another document, XBL2's spec:

   http://www.w3.org/TR/xbl/

This is just the CSS-specific side of it.

To give some background, originally I wanted the spec to be developed 
completely in the CSSWG, and the intention was just to continue the BECSS 
work with that spec, but Bert didn't want XBL in the CSSWG, so it went to 
the WAFWG instead. However, it still had CSS bits inside, and so the CSSWG 
asked that those bits be moved back to CSS. Hence, they landed in the 
BECSS draft, just as was originally intended for the whole thing. Reusing 
the draft is a convenient way of announcing that the group has decided on 
a new direction for the behavioural extensions work, without leaving 
dangling working drafts that people will bug us about.


> >> The original document was far more than just a first WG - Microsoft 
> >> implemented a feature called CSS Behaviors, and it's in fairly heavy 
> >> use today (for VML, SMIL applied to HTML, and a bunch of custom 
> >> controls as well).
> >
> >I don't think we really should count features that are used by the UA 
> >itself to implement other features as "heavy use".
> 
> You've heard of a product called Microsoft Office, I presume?

Microsoft Office uses VML and Microsoft's SMIL predecessor HTML+TIME, it 
doesn't use HTCs (except insofar as it needs to to use VML and HTML+TIME, 
but that's an implementation artefact, not what I would consider a "use").

In any case, Microsoft Office's proprietary extensions to Web standards 
that enable round-tripping to and from Microsoft's own products hardly 
count as examples of "heavy use", nor do they affect this situation -- 
authors are the ones that might hypothetically get confused by us naming 
this draft BECSS, not Microsoft's engineers.


> >> I'm just saying "behavior" is a loaded word, and we should either be 
> >> making this CSS feature either really inherit spiritually from the 
> >> previous one
> >
> >It does. BECSS and HTCs (Microsoft's behaviours) and Action Sheets 
> >(also in the original BECSS draft) inspired Mozilla's XBL, which was 
> >the base for sXBL, which is where XBL2 came from, which is what the new 
> >BECSS draft is primarily based around today.
> 
> That's not a strong enough reasoning to reuse precisely the same name.  
> Imagine if Netscape Navigator still had 80% market share today, and I 
> proposed a "Layer" element in HTML5.  I am merely trying to avoid 
> confusion.

The difference is that an actual <layer> element would (theoretically) 
actually affect existing implementations, whereas merely reusing the name 
"BECSS" doesn't affect anything from a technical perspective.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
Received on Tuesday, 23 October 2007 01:11:45 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:55 GMT