Re: 3 questions on validation

In article <3q2j7fFaufdvU3 (AT) uni-berlin (DOT) de>,
Gérard Talbot <newsblahgroup (AT) gtalbot (DOT) org> wrote:

http://www.gtalbot.org/Varia/Validat...lanations.html

Therefore, the best starting step to ensure a webpage appears
and works as expected for all modern web browsers and web rendering
software is to write HTML code in conformance to the specifications.

But validity does not imply conformance. Also, it is possibly to be
valid but still incompatible with the real world.

http://hsivonen.iki.fi/test/minimization.html

Generally speaking, a webpage using valid HTML code is smaller in
size (so downloading is faster) and is also rendered
considerably faster in modern browsers.

What's the basis of those (false) claims?

The W3C provides a public HTML validator service that
automatically checks a submitted web page against the formal
rules (semantic, syntax) of the HTML specification and then
report any error found.

The W3C validator makes no semantic checks whatsoever.

The clickable W3C HTML 4.01 button-image on webpages provides a
way for any vistor to verify that such webpages' HTML code is
properly coded.

http://www.cs.tut.fi/~jkorpela/html/...tion.html#icon

Passing HTML validation does not necessarly guarantee that the
webpage is an overall good webpage or a well designed one.

Nor does it guarantee conformance. It does not even guarantee that the
parse tree in the browser is the same as in the validator.

How about just linking to
http://www.cs.tut.fi/~jkorpela/html/validation.html ?

--
Henri Sivonen
hsivonen (AT) iki (DOT) fi
http://hsivonen.iki.fi/
Mozilla Web Author FAQ: http://mozilla.org/docs/web-developer/faq.html