[ANN] markup validator 0.8.3 released

[olivier Thereaux]

Dear all,

We released today a new version of the Markup Validator:
online http://validator.w3.org/
documentation http://validator.w3.org/docs/
download http://validator.w3.org/source/

0.8.3 is mostly a maintenance/bug fix version, but it also has a few
nice new features, especially for developers.

For more info on the change, see the new page:
http://validator.w3.org/whatsnew.html#t2008-08-08
...or this little writeup on the Q&A weblog:
http://www.w3.org/QA/2008/08/markup_validator_updated.html

Many thanks to those who have helped with this release, by reporting
bugs, suggesting fixes or documentation tweaks, or provided patches.


** Send Feedback - Get Involved **

The markup Validator is one of the many open source software projects
created and maintained at W3C: your participation in sending feedback
and ideas, patches and bug reports helps build better tools.

The mailing-list www-validator is the primary discussion channel for
users and contributors of the Markup Validator (and Link Checker).
Please make sure to search the archives before sending in a new idea
or request, as a courtesy to the many readers of the list.
http://lists.w3.org/Archives/Public/www-validator/

Bug reports and patches can be sent to the public W3C Bugzilla
database. Usual bug reporting guidelines apply - make sure your bug
is a bug, make sure it is not yet known, and give precise and
complete information to reproduce.
http://www.w3.org/Bugs/Public/

For an overview of feedback mechanisms for the validator, see:
http://validator.w3.org/feedback.html


Thank you,
olivier.
--
olivier Thereaux - W3C - http://www.w3.org/People/olivier/
W3C Open Source Software: http://www.w3.org/Status

[Rick Merrill]

Why can't the CSS validator "resolve" a relayed URL, such as  
www.batco.tv  ?

[Frank Ellermann]

olivier Thereaux wrote:
 
> We released today a new version of the Markup Validator

Great, thanks.
 
> For more info on the change, see the new page:
> http://validator.w3.org/whatsnew.html#t2008-08-08

Something with the added "historical HTML i18n" isn't as
expected:

| the Document Type (-//IETF//DTD HTML i18n//EN) is not in the
| validator's catalog

Did you use another public identifier ?

The "whatsnew" file says:

| Conformance: The validator now warns about incorrect
| public/system identifiers combinations

They are not really "incorrect", or are they ?  The
warning uses the adjective "inconsistent".  I still
wonder how that's supposed to work:

The validator knows some public identifiers (but not
all).  Where that is not the case the validator uses
the system identifier to find the DTD.

Users are in the same position wrt to their own local
tools, and their set of "known" public identifiers is
different from what the validator knows.

Shouldn't they be able to use system identifiers that
make sense from their local POV ?  I vaguely recall
that some system identifiers for W3C DTDs used to be
relative URLs.  IIRC that was the reason why I used
absolute URLs making sense from my POV, URL's in the
validator's catalog...  Which are not the "official"
system identifiers.

 Frank

>
> Many thanks to those who have helped with this release, by reporting
> bugs, suggesting fixes or documentation tweaks, or provided patches.
>
>
> ** Send Feedback - Get Involved **
>
> The markup Validator is one of the many open source software projects
> created and maintained at W3C: your participation in sending feedback
> and ideas, patches and bug reports helps build better tools.
>
> The mailing-list www-validator is the primary discussion channel for
> users and contributors of the Markup Validator (and Link Checker).
> Please make sure to search the archives before sending in a new idea
> or request, as a courtesy to the many readers of the list.
> http://lists.w3.org/Archives/Public/www-validator/
>
> Bug reports and patches can be sent to the public W3C Bugzilla
> database. Usual bug reporting guidelines apply - make sure your bug
> is a bug, make sure it is not yet known, and give precise and
> complete information to reproduce.
> http://www.w3.org/Bugs/Public/
>
> For an overview of feedback mechanisms for the validator, see:
> http://validator.w3.org/feedback.html
>
>
> Thank you,
> olivier.
> --
> olivier Thereaux - W3C - http://www.w3.org/People/olivier/
> W3C Open Source Software: http://www.w3.org/Status
>
>

[Shane McCarron]

Frank Ellermann wrote:
> The "whatsnew" file says:
>
> | Conformance: The validator now warns about incorrect
> | public/system identifiers combinations
>
> They are not really "incorrect", or are they ?  The
> warning uses the adjective "inconsistent".  I still
> wonder how that's supposed to work:
>  
I would have to agree with this.  The SYSTEM identifier is permitted to
change by definition.  If you look in the XHTML specs, for example, they
all say something like "There MUST be a DOCTYPE declaration in the
document prior to the root element. If present, the public identifier
included in the DOCTYPE declaration MUST reference the DTD found in ...
using its public identifier. The system identifier MAY be modified
appropriately."  I don't think having the validator warn about SYSTEM
identifiers that are different is a useful feature.  If the validator
wants to use its local cached versions of DTDs, what it should do is
check the SYSTEM identifier against its perception of correct.  If it
matches, great.  If it does not, dereference the one specified by the
document and use it.  Otherwise the validator is not doing what I (at
least) would expect.  I know that is not how an SGML catalog
traditionally is used, but in this case I think it is appropriate.  Or,
alternately, have a checkbox that means "use the DTD I specify".

And no, I don't know why I always capitalize SYSTEM.

--
Shane P. McCarron                          Phone: [phone] x120
Managing Director                            Fax: [phone]
ApTest Minnesota                            Inet: shane@...

[olivier Thereaux]

Hi Frank,

On 8-Aug-08, at 7:22 PM, Frank Ellermann wrote:
> Something with the added "historical HTML i18n" isn't as
> expected:
>
> | the Document Type (-//IETF//DTD HTML i18n//EN) is not in the
> | validator's catalog
>
> Did you use another public identifier ?

PubID      = -//IETF//DTD HTML i18n/EN
ouch, looks like a typo in
http://dev.w3.org/cvsweb/validator/htdocs/config/types.conf.diff?r1=1.35&r2=1.36&f=h

> The "whatsnew" file says:
>
> | Conformance: The validator now warns about incorrect
> | public/system identifiers combinations
>
> They are not really "incorrect", or are they ?  The
> warning uses the adjective "inconsistent".

Would you call this correct?

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN"
   "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">

That was precisely the kind of cases we found in the wild.

See:
http://www.w3.org/Bugs/Public/show_bug.cgi?id=4848

A lot of people edit the DOCTYPE by hand even when they don't have a  
clue what it is. If the validator tells them "your doctype doesn't  
look right" we may help them fix their error.

Sure, this may not be "right" in theory, but in practice, I'm sure  
throwing a warning to enforce that DOCTYPE declarations are consistent  
with what the specs say does more good than harm.

> Users are in the same position wrt to their own local
> tools, and their set of "known" public identifiers is
> different from what the validator knows.
>
> Shouldn't they be able to use system identifiers that
> make sense from their local POV ?

You are talking about a very small fraction of users who
* know their SGML by heart,
* know what they are doing when they edit a SI in a document,
*and are able to *understand* the warning from the validator, and  
*assess* whether it applies.

For the rest of us, the warning is useful.

[Frank Ellermann]

olivier Thereaux wrote:

>> Did you use another public identifier ?

> PubID      = -//IETF//DTD HTML i18n/EN
> ouch, looks like a typo

First fox for version 0.8.4 :-)

>>| Conformance: The validator now warns about incorrect
>>| public/system identifiers combinations

>> They are not really "incorrect", or are they ?  The
>> warning uses the adjective "inconsistent".
 
> Would you call this correct?
 
> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN"
>    "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
 
> That was precisely the kind of cases we found in the wild.

Ugh, I needed some time to get the 1 != 1.1 joke.
 
You convinced me months ago that a warning is good.

The warning says "inconsistent".  IMHO using the same
adjective "inconsistent" also in whatsnew.html would
be more consistent than writing "incorrect".

I don't use XHTML 1.1, but I'd know where to find a
flat DTD for an inconsistent (but hopefully correct)
system identifier for this beast, here:

http://validator.w3.org/sgml-lib/REC-xhtml11-20010531/xhtml11-flat.dtd

> http://www.w3.org/Bugs/Public/show_bug.cgi?id=4848

Terje's comment confirms that using adjectives like
"broken" or "incorrect" can be too harsh.  

The flat DTD above is hopefully not "broken", because
it is what the validator actually uses.  Even if folks
manage to use the "correct" system identifier.

This spooked me years ago, while the validator still
used the 1st edition of XHTML 1, and I wanted the 2nd
edition.  I forgot the details, but after that event
I used "inconsistent" validator.w3.org DTDs in system
identifiers where possible.

> in practice, I'm sure throwing a warning to enforce
> that DOCTYPE declarations are consistent with what
> the specs say does more good than harm.

Yes, I didn't intend to restart this discussion. The
warnings are okay.  Please s/incorrect/inconsistent/
in whatsnew.html for consistency :-)

> You are talking about a very small fraction of users who
> * know their SGML by heart,
> * know what they are doing when they edit a SI in a document,
> *and are able to *understand* the warning from the validator, and  
> *assess* whether it applies.

It was because I did *NOT* know what I'm doing when I
tried to "fix" the system identifiers years ago.  Only
much later I figured out that the validator ignores
the system id. for all DTDs it knows by its public id.,
and my wannabe "fix" was a complete waste of time.

But in a certain sense it was very "consistent", today
I'd just ignore the warnings.  Maybe I update my XHTML
template to be "correct" when I get a round tuit.

 Frank

[olivier Thereaux]

On 10-Aug-08, at 1:27 PM, Frank Ellermann wrote:
> olivier Thereaux wrote:
>> PubID      = -//IETF//DTD HTML i18n/EN
>> ouch, looks like a typo
>
> First fox for version 0.8.4 :-)

Indeed.

> The warning says "inconsistent".  IMHO using the same
> adjective "inconsistent" also in whatsnew.html would
> be more consistent than writing "incorrect".

Aha, agreed.

Both changes done in CVS and live on the service.
They will "officially" be in the release tarballs for the next version.

Thanks!
--
olivier

[Frank Ellermann]

olivier Thereaux wrote:

> Both changes done in CVS and live on the service.

Thanks.  May it help future software archaeolgists
when they try to reconstruct how "ASCII or else"
ended up as "UTF-8 everywhere" in 40+ years... :-)

 Frank