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

Re: getting rid of callers

From: L. David Baron <dbaron@dbaron.org>
Date: Fri, 29 Oct 2010 07:12:52 +0200
To: Ian Hickson <ian@hixie.ch>
Cc: Jonas Sicking <jonas@sicking.cc>, Cameron McCormack <cam@mcc.id.au>, public-html@w3.org
Message-ID: <20101029051252.GA9055@pickering.dbaron.org>
On Wednesday 2010-10-13 08:09 +0000, Ian Hickson wrote:
> On Tue, 12 Oct 2010, Jonas Sicking wrote:
> > >> For HTMLFormElement, it seems only IE implements them.  Are they 
> > >> really necessary to have in the spec?
> > >
> > > The real question is, is Microsoft willing to remove support for this 
> > > feature? (in all modes, not just in some DOCTYPE-triggered ghetto)
> > >
> > > If not, then the simplest way of getting interop would be for everyone 
> > > to just implement these features. They're not that much of a burden, 
> > > surely?
> > 
> > This is clearly not the only qualification for if a feature should stay. 
> > Or are you arguing we should add VML and vbscript to HTML5 as well?
> 
> There's a world of difference between supporting making an object callable 
> and implementing an entire scripting language. I wouldn't expect the 
> burden of implementing an entire scripting language to be a light thing.
> 
> Personally I just want interop. I don't really mind how we get there. I 
> also acknowledge that the journey is long. We have to do it piecemeal.

I don't think we should standardize every browser maker's failed
(due to lack of adoption) experiments at feature additions just
because the browser maker won't remove them.

There are good reasons to value simplicity in the Web platform:

We want it to be easier rather than harder to create new
implementations of the Web platform, since new implementations are a
good source of innovation and major improvements.  (Consider
particularly new implementations that are fundamentally different
from current ones, for example in being better able to exploit
hardware parallelism, or being written in much safer languages.)

We want Web content to be durable.  Hundreds of years from now,
historians will want to read content written today, even if they
can't run today's software.  While standardizing everything that
ever was may have short-term advantages in durability, I think in
the long term it's a loss due to the complexity of creating new
implementations.

We also want the Web platform to be understandable to authors.
Additional complexity makes this harder.  Many ways to do the same
thing is often bad:  it makes the author have to choose one of them,
and it makes it harder for one author to read code written by
another.  (Consider perl vs. python.)

-David

-- 
L. David Baron                                 http://dbaron.org/
Mozilla Corporation                       http://www.mozilla.com/
Received on Friday, 29 October 2010 18:08:13 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:17:16 GMT