Re: The state of our web site and standards

[Daniel Veillard <veillard redhat com>]

On Thu, Apr 12, 2001 at 03:10:31PM +0300, Ville Pätsi wrote:
> These 2.6K bytes were from Telsa Gwynne,
> > Agreed, Bobby is picky. So is htmltidy (which demands summary attributes
> > for any table). 
> 
> This raises an another problem. I frequently use the program "tidy" to verify
> my html. This program is created by a person from the w3c. It validates the 
> code of my homepage as complete html 4.0. ( http://canvas.gnome.org/~drc/ )
> But when I run it trough http://validator.w3.org/ , it gives a bunch of errors.
> Now which program is right? How are we supposed to produce valid html if the
> checker programs don't validate their results between each other?
>   Before going into a discussion about forcing people to use validators, we 
> should choose which one to support and use. The best way propably would be 
> that the w3 create a piece of transportable code that would be used by all
> validators.

  I think I can try to at least answer:
    - tidy is what it's name says, a cleaning program, it doesn't
      have a full power of a full validation processor
    - validator.w3.org on the other hand is absed on James Clark
      SGML processing tools which are considered a reference implementation
      of SGML. for the SGML familly of HTML languages (i.e. not
      XHTML ones) you can consider the result of the validator running
      with one of the official DOCTYPE nearly normative. It's not
      up to par with XML or when using non ISO Latin encodings.

Basically the tools were designed with very different approaches in mind,
I don't think tidy was designed with really validation in mind, rather
correcting common mistakes, cleaning up proprietary stuff and allowing
to convert ot XHTML-1.0,

Daniel

-- 
Daniel Veillard      | Red Hat Network http://redhat.com/products/network/
veillard redhat com  | libxml Gnome XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/