Fwd: Bug report - no validation of URIs, not at even the most basic level

This message was sent to the admin address, should have gone to the  
list...

Begin forwarded message:

> From: "Cecil Ward" <cecil@cecilward.com>
> Date: August 7, 2007 04:26:32 JST
> To: <www-validator-request@w3.org>
> Subject: RE: Bug report - no validation of URIs, not at even the  
> most basic level
>
> About the Unicorn / URI validator component idea -
>
> A few initial thoughts.
>
> (1) Outside help: It would be worth investigating options for the  
> avoidance
> of wheel-reinvention. Who knows, some such high-quality code for  
> this job
> possibly already exists outside of the W3C. For example, reading  
> the MSDN IE
> blog, I read that Chris Wilson's team has put in a lot of work into  
> the
> development of a parser for URIs as part of a big security push in  
> the IE7
> project. It would be worth asking around to see if any member  
> organization
> might already have some suitable code as an internal test tool,  
> which could
> be re-used either in terms of code reuse, design reuse or even  
> stealing a
> human for a short period. In fact, perhaps MS even might do the  
> work if
> asked very nicely? For some good PR. Or just because they're good  
> people.
>
> (2) Modularization: It would certainly be worth modularizing such a  
> project
> in order to shorten delivery time-to-first-benefit as well, as  
> implementing
> at least the http: scheme first, trying to do something to handle the
> use-cases of relative URIs and handling of fragment-identifier-only  
> URIs
> would give the maximum benefit first by covering a lot of common  
> use-cases.
> Even if the validators had to say "unable to check" a lot in the  
> interim,
> then that would be fine for a start.
>
> (3) Disclosure in the short term: It would be good if various web  
> pages were
> amended now, so that the existing validators owned up to this  
> limitation
> right away, as a point of information for users. The tools might  
> simply
> state "URIs not checked" in the interim at a warning- or
> informational-level. (Precedent: We are all already familiar with  
> the way
> that many accessibility checking tools happily simply issue  
> warnings of
> "rule x not checked" for must-be-checked-manually issues.) The QA- 
> Dev tools
> summary page http://www.w3.org/QA/Tools/ could also acknowledge  
> this gap
> too, as an advertisement.
>
> Best,
>
> Cecil Ward.
>
>

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

Received on Tuesday, 7 August 2007 01:33:12 UTC