W3C home > Mailing lists > Public > www-validator@w3.org > October 2004

Re: Standards

From: Tim Jackson <lists@timj.co.uk>
Date: Thu, 7 Oct 2004 10:14:26 +0100
To: "Abyss" <info@abyss.ws>
Cc: <www-validator@w3.org>
Message-ID: <E1CFULy-0005cC-JX@virt1.clust1.firecluster.net>

Hi Abyss, on Thu, 7 Oct 2004 18:42:14 +1000 you wrote:
 
> Then what is the point of having standards if the same code does not
> work in different browsers?

I think you're missing the point of both the Web in general and validation
in particular. The Web is more than making sites "look good" in a couple
of specific visual browsers. Here are a few thoughts to bear in mind:

- user agents ("browsers") don't have to be visual. A few examples of ways
people might access the web include on a PDA, mobile phone, via an in-car
console or many other weird and wonderful ways that neither you or I have
considered. That's the beauty of the Web: it's open, and people are free
to use it in the way that suits them rather than the person developing the
site, as long as the person developing the site does it properly.

- user agents don't have to be either IE or "Netscape" (now mostly
Mozilla)

- (X)HTML/CSS is not a pixel-perfect page layout language.

- Validation does not assure that a page is going to look the same in
every browser. In fact, that's impossible. If you look at a site in Lynx
(a text-mode browser), for example, then clearly it's not going to look
the same as in a modern visual browser regardless of whether it's valid or
not. If you're visually impaired and using a site with a screen reader or
other accessibility technology then it's definitely not going to appear
the same.

> it defeats the point.

No, it doesn't - it's just that you're misunderstanding "the point". The
point of having valid HTML is that it's an important part (not the only
factor by any means, but an important one) in ensuring that the site can
be read and understood by any compliant user agent ("browser").

However, in actual fact, developing standards-based sites *does* help the
consistent presentation. From a purely visual point of view, you'll find
that well-formed, standards compliant sites generally look very similar in
equivalent installations of popular modern browsers such as Mozilla, Opera
and Konqueror. Internet Exploder too, if you're willing to sacrifice most
the functionality of CSS2 and stick to a subset of its capabilities.

If you have a non-valid page then effectively all bets are off because
you're relying on undefined error-correction within browsers.

> ok browsers have bugs - understandable, but which browser is correct? IE
> or NS?

As a general rule, Mozilla and Opera are by far the best at rendering
things according to standards. They are both pretty good. IE sucks
extremely badly and has some huge bugs which only serve to demonstrate the
utter incompetence and ignorance of the people who developed it. Just to
make life even harder for you, the ways in which it sucks vary
considerably between versions and even between platforms (IE5 on Mac is
substantially different to IE5 on Windows, for example. In fact, they're
pretty much different browsers.)

Netscape 4, along with some other older browsers, is pretty much a lost
cause, at least if you're trying to do anything much involving CSS. But
that's cool, you can still make things work in it - they'll just be plain
with basic structural formatting only.

> or should i just stick the goldern rule of  - "program for the browser
> that is most widely available"?

No, that's silly. Being pragmatic, you may well want to verify that sites
work well in popular visual browsers. However, with a bit of effort, some
limitations and the occasional workaround, it's entirely possible to
develop sites which are both standards-compliant *and* work well in all
popular browsers, including that crap one with a big blue E which so many
people have an unhealthy addiction to.


Tim
Received on Thursday, 7 October 2004 09:14:31 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 25 April 2012 12:14:15 GMT