W3C home > Mailing lists > Public > www-html@w3.org > April 2004

Re: complexity

From: <olafBuddenhagen@web.de>
Date: Sat, 10 Apr 2004 11:11:42 +0200
To: www-html@w3.org
Message-ID: <20040410091142.GA826@sky.local>


On Fri, Apr 09, 2004 at 07:37:11PM +0000, Ian Hickson wrote:

> >> I'm confused; you want the working group to _not_ address author
> >> wishes?
> >
> > I want the working group to carefully distinguish between reasonable
> > things, that give a considerable benefit at an acceptable cost, and
> > silly ideas that only add unnecessary complexity.
> We do. Maybe our opinion of what is silly is different, though. :-)


> >> If two UAs can implement it, why would a third not be able to?
> >
> > Because of cost. Not every browser vendor has such an enormous
> > manpower.
> Mozilla and Opera both have _extremely_ small development teams. If I
> recall correctly the core rendering engine team at Opera is no more
> than 6 people, and it is about the same number of people for Mozilla.
> And this covers everything: SVG, CSS, HTML, XML, plugins, events, DOM,
> forms, HTTP, cookies, javascript, etc.

Speaking of "core teams" is misleading. It doesn't change the fact that
you need several dozens of developers and five years to create a
full-featured web browser that is terribly slow, implements only a
fraction of W3C standards, and is terribly buggy with others. Multiply
this by a factor to get a *fast* and bug-free implementation, another
factor to get support for all existing and upcoming standards, then
something more to have it finished somewhat faster, don't forget Brook's
law, and you'll come to the conclusion that you need several hundred
developers to create a really good browser in a reasonable amount of

Only now making this calculation I realize *how* bad things really are

> Anyway, I would recommend being more specific in your criticisms.

There is no point in being more specific in DOM or CSS criticism on a
HTML list. The whole point of this thread was to underline the
*fundametal* problems.

As for XHTML 2, I've already mentioned the things that give me a
headache: XForms and XInclude. Oh, and <nl> of course, if I really
wanted to implement it the way it is described in the standard.
Otherwise, the current draft is really nice in terms of implementation

Received on Saturday, 10 April 2004 08:00:11 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:06:08 UTC