Errors in MarkUp Validator, Release 0.6.5

(first message wasn't wrapped - sorry - this should be better)

I have been using both the MarkUp validator and the CSS Validator
for more than a year - most of the time with great satisfaction.
But the recent upgrade to 0.6.5 is disappointing I'm sorry to say.

I'm glad that the validation service is undergoing development, and
I think that some of the announced improvements are very good - but
the problem is that some of these improvements doesn't seem to
work. Another problem is that important options on the revalidate
page are now gone.

The last thing first. Consider the following page:
http://www.gyros.dk/nodtd.html

This page lacks a doctype declaration and a charset encoding, but
is otherwise valid xhtml. With the old validator I would get an
error message telling me that the validator could not find at
charset definition, and the validator would then let me specify a
charset override. The new validator gives another error message -
telling me that the document cannot be validated because it
contains bytes that can't be interpreted as utf-8. I might be o.k.
that the validator automatically guesses "utf-8" when a charset
definition is missing, but it is a major problem that the validator
fails to give me an option to revalidate with charset override. I
get "Revalidate with options" and a small list of checkbuttons, but
the must important revalidate options - doctype and charset options
are hidden. Why is that?

Now concerning the error.
On the detailed validation page -
http://validator.w3.org/detailed.html - it is now possible to
specify "fallback instead of override". The options are explained
in the help page. But the fallback options does not work as the
help text explains.

Try to validate the following page without specifying any extra
options: http://gyros.dk/usenet/

The page should validate without problems.

Now try specifying fallback to HTML 4.01 Transitional. The
validation should - if I understand the help text correctly - still
pass, since the fallback doctype should only be used if the page
fails to specify one. But the result is the following:

	No DOCTYPE Found! Falling Back to HTML 4.01 Transitional

And if one puts a mark in "show source" - the original XHTML
doctype is magically gone!

Another error. Choose another doctype as fallback - i.e. XHTML 1.0
Transitional - and validate. Once again you get:

	No DOCTYPE Found! Falling Back to HTML 4.01 Transitional

So not only has the validator "eaten" the original doctype, it also
fails to use the specified fallback type. No matter what doctype I
choose as fallback, the validator uses HTML 4.01 Transitional.

The encoding option works a bit better - but still not correct. If
I choose fallback to utf-8, the validation uses an override to utf-
8.


The new possibilities with the validator would be very welcome if
they worked - but for now I would be happier if I could use the old
validator.

Regards
Jens Gyldenkaerne Clausen

Received on Wednesday, 12 May 2004 13:15:03 UTC