Re: Tidy bug (malformed URI reference)

At 05:15 PM 12/5/2002 +0100, Bjoern Hoehrmann wrote:
>* Charles Reitzel wrote:
> >Unfortunately, this is a case of "damned if you do, damned if you
> >don't".
>
>Indeed, so authors should avoid using spaces or non-ASCII characters in
>URIs, no matter what scheme.

"Don't do that" is hardly consistent with Tidy's goal of fixing up markup, 
is it?  We usually only resort to that answer when faced with a 
functionality conflict or sheer difficulty.  We have neither in this case.


> >FWIW, I couldn't find Javascript links <a href="javascript: ...">foo</a>
> >described anywhere in the HTML spec.
>
>The scheme is not registered and due to usability and accessability
>issues, no one should use it, especially not the HTML recommendation.

Well, Tidy supports a number of HTML features that are not defined by any 
W3C Recommendation.  So, fine, it isn't in the rec. However it is already 
in use by thousands, if not millions, of pages.  The typical use is for 
opening a link in a new window with a certain size and position.  It is a 
bit late to say, "no one should use it."


> > So we may be in a gray area of
> > compliance with the _Recommendation_.
>
>No, there is a normative reference to RFC 2396 which forbids spaces in
>URI references, and there is
>
>   http://www.w3.org/TR/html4/appendix/notes#h-B.2.1
>   http://www.w3.org/International/O-URL-and-ident.h

References to RFCs are only as normative as the Recommendation itself - 
which isn't.  It is a recommendation.  This fact is clearly demonstrated by 
the lack of compliance with all (X)HTML Recommendations by all known 
browsers and other clients.  In fact, Tidy exists entirely because of the 
gaps between the recs and working implementations.

Again, Javascript URLs are a de facto standard that works reasonably well 
on  all browsers and other markup processors.


>Tidy does what is written in the specification,
>such behaivour is not a bug and won't be changed.

That is your opinion.  I hope you agree that what Tidy does and what will 
or will not be changed is ultimately a consensus decision by the group.

take it easy,
Charlie

Received on Thursday, 5 December 2002 12:45:30 UTC