W3C home > Mailing lists > Public > public-html@w3.org > March 2010

Re: Bug 7034

From: James Graham <jgraham@opera.com>
Date: Fri, 26 Mar 2010 17:30:38 +0100
Message-ID: <4BACE12E.6040206@opera.com>
To: Sam Ruby <rubys@intertwingly.net>
CC: Henri Sivonen <hsivonen@iki.fi>, HTMLwg WG <public-html@w3.org>
On 03/26/2010 04:53 PM, Sam Ruby wrote:
> On 03/26/2010 10:20 AM, Shelley Powers wrote:
>>>
>>> Apparently, you got some blowback from the CSS-orthodoxy.  We also 
>>> have an
>>> accessibility-orthodoxy, an XML-orthodoxy and possibly quite a few
>>> additional orthodoxies to deal with, each of which require that we 
>>> worship
>>> in their various churches.
>>
>> I'm hoping that you're being facetious, because what you've done is
>> pretty much described a majority  portion of the user base for the
>> HTML5 spec -- and Apple's, Mozilla's, Google's, IBM's, Opera's, and
>> Microsoft's customers. And in a way that struck me as being
>> dismissive.
>>
>> You all are enjoying this quite long semi-permathread, and that's
>> cool. But when you referring to significant proportions of this
>> group's user base as "standardistas", and  seemingly dismiss the
>> concerns of large numbers of people in some form of tech elitist
>> frenzy, maybe you need to step back and re-establish your perspectives
>> a bit.
>
> Orthodoxy is NOT a bad thing.  When I was growing up, the rule was 
> fish on Fridays.  I have friends that avoid pork, and ones that avoid 
> all meat.

I think that framing this discussion in terms of religion is unlikely to 
have the effect of drawing it rapidly to an amicable conclusion.

> What we have here is a number of groups here who take the position 
> that my rules rock, your rules suck.

I haven't really seen this attitude that you describe. In particular the 
majority of the recent discussion I have seen has suggested that the 
participants are aware of the complex nature of the problem and are 
prepared to shift their views when new information or arguments are 
presented.

> I want to take a consistent position.  One where we are inclusive of 
> everybody's rules, or at least one where treat each constituency 
> consistently.

So essentially you advocate normative relativism; that is a fair 
characterisation of your position would be that it it is not our job to 
perform any evaluation of the merits of a given ruleset, but simply to 
let anyone pick whatever rules they please?

> I want each set of rules to be captured.  I want each set of rules to 
> be strongly encouraged. 

I don't see how that is possible given the possibility of mutually 
contradictory preferences (consider: "you should use as little markup as 
possible" vs "you should explicitly include all tags and end tags, and 
quote all attribute values").

> Where I draw the line is that I don't want people to be able to say 
> that google.com is incorrectly using the text/html mime type simply 
> because they don't escape their ampersands.

I think that we should be prepared to look at the available data, have 
reasoned discussions based on that data, and have a process for 
converging on a set of rules that meets whatever we determine the 
essential criteria to be (for example, accessibility, i18n, helping 
authors avoid rough edges in the language that cause undue pain), even 
when not everyone agrees on the perfect route forward. I see the 
alternative approach as an abdication of our responsibilities to authors 
and end users. For example, in the case of ampersands, we might, through 
the process of actually investigating current usage, find it is possible 
to change the parsing so that entities are not expanded in attribute 
values, thus making it possible to simplify the language and make 
unescaped ampersands in attributes legal. By simply punting on the 
problem we would never make that discovery.

Another example is <b><i></b></i>; by one criterion that has been thrown 
around ("UA-level interoperability"), that would not be a conformance 
error. Therefore an author getting unexpected behavior because of a 
simple-to-make error would not only have to think to use a validator, 
but also to load the correct rulesets in order to find the problem. I 
don't think we should expect more from authors just because it would 
mean fewer tough decisions for us.
Received on Friday, 26 March 2010 16:31:40 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 29 October 2015 10:16:00 UTC