Re: href="http:something.html", and advocady

On Thu, 24 Jan 2002, Stephane Gourichon wrote:

> I've noticed that some sites (the latest I found is
> http://www.lri.fr/~aze/page_c/aide_c/c_gauche.html ) have an unusual way
> of writing URIs, something like HREF="http:union.html"...
>
> Those are cited in rfc2396, for example, but "should be avoided".
> (See excerpts below this mail.)
>
> Netscape 4.x, Lynx 2.8.1, accept this scheme.
>
> Links 0.96 is unaware, and interprets this way:
> http://www.lri.fr/~aze/page_c/aide_c/http:union.html
>
> Galeon 0.12.1 doesn't like it (not checked exactly, looks like it does a
> ns lookup on "union.html").
>
> I don't know if the w3c validator, or the WDG validator check the
> conformance of URI's in documents

No, they don't.

> Also, do you think that in such case I should e-mail the author of such
> pages, gently complaining that accessing his documents is a bit
> difficult and citing the validator's URL ?

It shouldn't hurt, and it could help.

> This one: http://www.htmlhelp.com/tools/validator/reasons.html is bad,
> because it it based on particular versions of browsers, and is bound to
> be outdated.

The same could be said of some authors' HTML--and that's the point of that
page.  When each particular version of Netscape was released, some pages
that used invalid HTML suddenly didn't work properly in the new version
even though they worked fine in the old.  Authors who had validated their
HTML had no problem, but many authors had to go back and fix their HTML
with each major release of Netscape.

So the point of reasons.html isn't that Netscape x.y doesn't handle some
invalid HTML.  The point is that when Netscape x.y was released, some
invalid pages suddenly became broken to a large portion of Web users.
Testing pages on the most popular browser of the day wasn't enough to
survive new browser releases.

(Many people seem to be missing the point of reasons.html.  If anyone has
suggestions for making it more clear, please let me know.)

-- 
Liam Quinn

Received on Thursday, 24 January 2002 12:15:09 UTC