W3C home > Mailing lists > Public > www-validator@w3.org > June 2007

Re: Microsoft writing libs that produce bad markup (again) (was www.validator.ca / www.validateur.ca)

From: David Dorward <david@dorward.me.uk>
Date: Tue, 12 Jun 2007 23:27:04 +0100
To: Greg Sabin <movingpictures4u@yahoo.com>
Cc: Nick Kew <nick@webthing.com>, www-validator@w3.org, marc@proze.net
Message-ID: <20070612222704.GA19476@us-lot.org>

On Tue, Jun 12, 2007 at 12:22:52PM -0700, Greg Sabin wrote:

> I was not able to get W3.org to make the validator work with
> Microsoft office live web sites.

If I remember rightly, the series of events was something like this:

* W3C produce a specification for HTML

* W3C and volunteers produce a tool to test against against various
specifications.

* Microsoft produce a code library that generates HTML.

* You use that library to generate a webpage.

* You test that webpage against the aforementioned HTML specification,
and it doesn't conform.

I can't see how anyone would be impressed by Microsoft's contribution
to that chain of events, or how the people writing the testing tool
could do anything about it. The only fixes that I can see are:

(a) Microsoft fixing their libraries

(b) Someone writing a specification for the code generated by those
libraries, this should be Microsoft, since reverse engineering a black
box is unlikely to produce as good a result as the authors of a piece
of code documenting it themselves.

(c) Not using the libraries that generate the bad code

Option b, in my opinion would be the worst such option since, as far
as I know, the deviation from standard HTML do not provide any
benefits (so this would require all tools that consume HTML to be
updated or continue to depend on error recovery routines in exchange
for no additional features or solved problems).

> The validator is actually Vary limited in what it can and can not do

It is fairly limited, in the same way that a pencil is fairly
limited. It does a job and it does it well.

> , and when I asked I was told in no short terms that I should make
> my web site conform to the validator , rather then have the
> validator be able to read multiple formats on a single web site.

I think the problem was that you were using an undocumented
format. How can you test for conformance to something without knowing
what the 'something' is ?

> I have noticed many vary good looking web sites can not pass the
> W3.org test , possibly W3.org is only interested in diction and
> grammar verses real life applications , free thinking and or
> thinking out side of the box.

When one thinks outside the box, it is usually to provide some
tangible benefit. What benefits over standard HTML do Microsoft's
non-standard changes provide? If there is no benefit, it is simply
unnecessary complication that forces user agents to compensate for the
deviations from the specification (making them larger, slower and less
stable).

> It would seem the people at W3.org don't like Microsoft or their
> programming vary much.

The quality of their generated HTML code is less then good, but this
issue is not so much about not liking Microsoft as it is about a group
of mostly volunteers refusing to invest a good deal of their time into
cleaning up Microsoft's mess for them. Correction, not cleaning it up
- sweeping it under the carpet.

-- 
David Dorward
http://dorward.me.uk/
http://blog.dorward.me.uk/
Received on Tuesday, 12 June 2007 22:27:21 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 25 April 2012 12:14:24 GMT