W3C home > Mailing lists > Public > www-validator@w3.org > May 2008

Re: web site programming.

From: Rui del-Negro <w3validator@dvd-hq.info>
Date: Sat, 03 May 2008 19:45:26 +0100
To: "Rick Merrill" <rickmerrill@comcast.net>
Cc: www-validator@w3.org
Message-ID: <op.ualc10cvjf0k3w@bigbang>

>>> And the Big League web sites - how about validating
>>> http://www.google.com/
>>> ???????????????!
>>>
>>> http://www.yahoo.com/
>>> ?!!!!!!!!!!!!!!!!
>>
>> How about it? The W3 isn't the HTML conformance police. I'm sure that  
>> if the people coding those sites want to validate their code, they will  
>> be able to find the validator (ex., by searching on... oh, I don't  
>> know, Google or Yahoo ;).
>
> what I am sayingis to create a COMPETITION taht the public can see.

The "public" has no idea what a validator is, or what it means for a page  
to be "valid". They don't understand that pages are not just processed by  
browsers, or that something that can be rendered by their browser might  
not work on a different one, or might cause a search bot to choke (as this  
thread shows, some people can't even tell the diffence between creating  
HTML markup and "web site programming"). And, even if average people  
cared, the chances of them locating the W3 validator and looking for some  
statistics there before deciding which HTML authoring software to buy is  
extremely slim.

There is no way to guarantee that a given page, tested on the validator,  
was completely generated by a specific package. So besides not being very  
useful, those statistics would be unreliable. Maybe the error was in a  
part the user had hand-coded, so the validator would think the package  
produced an error, when it hadn't. Or maybe the user had fixed some errors  
by hand, so the validator would think the package produced compliant  
mark-up when it didn't. Or maybe the user deleted the "generator" tag, so  
the validator wouldn't have a clue what software the page was created by.

Maybe something similar to the "ACID Test" could be created for HTML  
authoring applications, but that's really a separate project from a  
validator. It has to be tested in very controlled conditions, to guarantee  
that every aspect of markup generation is tested, and that code from  
different sources isn't mixed.

It's a job for software reviewers (on magazines, IT websites, webmaster  
forums, etc.), or for groups like webstandards.org, not for an automated  
conformance-checking tool.

RMN
~~~
Received on Saturday, 3 May 2008 18:46:23 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 25 April 2012 12:14:29 GMT