W3C home > Mailing lists > Public > www-validator@w3.org > December 2002

Re: Todo list

From: Frank Ellermann <Frank.Ellermann@t-online.de>
Date: Mon, 02 Dec 2002 22:32:23 +0100
Message-ID: <3DEBD167.1B6F@xyzzy.claranet.de>
To: Terje Bless <link@pobox.com>
Cc: W3C Validator <www-validator@w3.org>

Terje Bless wrote:

 [visible with any browser - NOT] 
> The links will probably not work in Netscape 4.x either

We totally agree that Netscape 4.x is as important as HTML 3.0,
so don't worry about it.  But Netscape 3 is a simple and robust
browser, not much bigger than Lynx & Co. on my "warped" system.

> It is unfortunate that this results in broken internal links
> for you of course, but we have to draw the line somewhere.

The line is "add comma tablin to get linearized tables", and it
(still ?) works.  It's IMHO inconsistent to validate something
like Wilbur or XTHML transitional, and then present the results
in a form not working with old browsers.

For a similar effect test the "diff-marked" version of e.g.
XHTML 1.0 2nd ed., there are no diff-marks at all with my old
browsers.  The essential content (the diffs) is hidden behind
HTML 4.01 constructs like <ins> and <del> resp. style sheets.

After comparing the HTML-sources I came to the conclusion, that
there's nothing new except from Appendix C.16 about &apos; in
the 2nd edition, and only weeks later I found that there are
many more changes in the DTD.  Only because somebody decreed,
that <del>...</del> is "better" than <s>...</s>. :-(

This is IMHO deliberately breaking old software for no visible
(or audible etc.) advantage at all.  It's possible to create
valid documents visible with more browsers (adding ",tablin" in
fact _all_ browsers), so why not do it ?  If WDG and Amazon can
do it, why should W3C fail ?

Just for fun I've even created a valid document using colors -
something I won't use in any other document - perfectly visible
with all browsers I can test (incl. a really old 1996 IBM OS/2
WebExplorer 1.1d, the latest version of Lynx, and "Bobby AAA"),
see <URL:http://frank.ellermann.bei.t-online.de/colour.htm>.

If it's possible with weird colors, then simple W3C validation
results should work with (almost) all browsers.  Maybe you need
something like a "XHTML classic DTD", tuned for compatibility
with most media and old browsers.  Unfortunately I still have
no idea how to write DTDs, otherwise I'd already created my own
mixture of XHTML Basic + Legacy + Names + (shudder ;-) <embed>.

 [Revalidation of valid documents]
> The next version will give you an option to show "Verbose"
> output, which should do what you want.

Thanks, although now that I know to simply add /detailed.html
to the URL I have my workaround,  But other users might still
miss the detailed interface.  And please add new doctypes like
"XHTML+1.1+plus+MathML+2.0" to the list of supported options.
 
 ["show source" parameter]
> For invalid pages there is currently no way to turn off the
> Show Source option. This may change in a future version. I've
> logged this request as Bug #97 in our tracking system.

Thanks again.  Of course I know that e.g. 2246 lines in a page
like <URL:http://frank.elermann.bei.t-online.de/ibm850.htm> are
far beyond any reasonable limit, but a simple warning would be
enough, I don't need my crap echoed back to me over a 56K line
to recognoize the problem with its size.  BTW, this document is
intentionally invalid, because I wanted to test some "unknown"
(new MathML) entities and the illegal &#128; upto &#159; stuff,
so I can't simply correct my errors to bypass your forced ss=1.
  
An option to sort errors by class instead of line numbers could
be handy for intentional errors (?), but probably I can also do
this on my side using your output=xml option.  

                           Bye, Frank
Received on Monday, 2 December 2002 17:59:36 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 25 April 2012 12:14:05 GMT