W3C home > Mailing lists > Public > www-validator@w3.org > October 2000

Re: Should Relative URL in format 'http:g/' validate

From: Terje Bless <link@tss.no>
Date: Sat, 28 Oct 2000 01:40:51 +0200
To: W3C Validator <www-validator@w3.org>
cc: Bjoern Hoehrmann <derhoermi@gmx.net>
Message-ID: <20001028014414-r01010600-84d6f26b@10.0.0.2>
[ Aaaargh! I *hate* Supercite[0] and it's ilk! Can't y'all just quote ]
[ in the accepted standard way for Internet email. *Must* you try to  ]
[ get clever? :-)                                                     ]

On 25.10.00 at 23:07, Bjoern Hoehrmann <derhoermi@gmx.net> wrote:

>"Terje Bless" <link@tss.no> wrote:
>
>>On 25.10.00 at 20:47, Bjoern Hoehrmann <derhoermi@gmx.net> wrote:
>>
>>>URI syntax checking and link checking is far beyhond the scope of what a
>>>Validator should do.
>>
>>Not really. The attributes in question have a specific data type which is
>>defined to be a valid URI according RFC"whatever". Checking that does
>>conform to those rules is not out of scope for the validator. More to the
>>point though, it's not out of scope as an extra feature that can be
>>enabled with a switch on the invoking form.
>>
>>The same goes for checking links.
>
>Validation is nothing more or less than checking some document against its
>associated DTD. A Validator should only do this validation. Everything
>that goes beyhond Validation is not a task that the validator should
>perform. The value of any href attribute in HTML 4.01 is defined as CDATA,
>the validator should look whether the value conforms to the given SGML/XML
>rules or not. If a href attribute value is valid CDATA the validator must
>not complain about that. If we extend the functionality of the W3C
>validator it is no longer a Validator.

"Let he who is without sin in his heart throw the first stone."

Are /you/ ideologically pure? Are you now, or were you ever, a member of
the Communist Party? Sorry I got carried away... :-)


The question of what is and isn't a "Validator" is really a matter of how
"ideologically pure" you are. In the strictest sense, yes, an "SGML
Validator" should check against the DTD and _nothing_ else. It shouldn't,
for instance, generate an outline of a document based on the Hn tags. It
shouldn't display a Parse Tree. It shouldn't report metadata pertaining to
HTTP. Or W3C Common Names for a particular DTD. In fact, it should be
_exactly_ what nsgmls is and nothing more.

However, as a matter of expediency and practicality, it's convenient and
useful to provide some extra services on top of strict SGML Validation.

Now /I/ believe that checking URI syntax and reachability falls within the
acceptable limit of what extra cruft you can bolt on to an SGML Validator
without diluting it's value. What we would be doing is checking that the
attribute value conforms to the rules set forth in the prose of the
reccomendation and that the URI is reachable. Since we plan to add
recursive checking anyway, this is a natural extension; we'd be doing the
work in any case so it's only natural to expose the information to the user
as well.


>I suggest reading Alan Flavell's
>http://ppewww.ph.gla.ac.uk/~flavell/www/validate.html

Alan is a very knowledgeable and eloquent person, but he does not have the
One True interpretation of all things WWW! While I usually tend to take his
opinion as canon, I must respectfully beg to disagree about the
interpretation on this one.


>A checker that performs further checking of (X)HTML is very useful, but a
>checker is no validator.

Non Sequitur! While a checker is not necessarily a validator, it can be,
and often is, a superset of a validator. The issue isn't whether or not it
provides extra services beyond pure validation, but rather whether it does
true validation at all.

Neither the W3C nor the WDG -- or even SP for that matter! -- are "true"
strict SGML Validators.


>To implement these features into the W3C Validator is not that easy.
>Currently it wraps the output of nsgmls into a HTML document. A checker
>needs another parser than nsgmls.

:-)

I'm still waiting for the documentation patches you promised me Björn...

Leave implementing it up to me and then you can have all the pleasure of
documenting it as the vile, evil, non-conformant, "Death of the .net
predicted. .mp3 at 11", mis-feature that it is. :-)


>I don't really believe that it is that worthly to create such a service.
>XHTML in combination with XML Schema could do a better job.

YM "XLink", surely?


Oh, and I keep hearing how wonderful everything is going to be when Schemas
arrive and how horrible DTDs are, but I've yet to see one comprehensible
argument for _why_ DTDs are so evil or _how_ XML Schema will bring about
the Second Coming. Explanations _gratefully_ accepted!








[0] - The spellchecker wanted to replace "Supercite" with "Supersex".
      What /are/ you smoking over there at BBSW? :-)


-- 
As a cat owner, I know this for a fact...
Nothing says "I love you" like a decapitated gopher on your front porch.
Received on Friday, 27 October 2000 19:44:24 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 25 April 2012 12:13:54 GMT