| That's a complete waste of bytes, IMHO. One can use the User-agent
| string to index a list describing compliance;
This is true - the client authors are not going know what bugs their
software has in order to say so in a Agent-Compliancy: line. Theres
quite a few other things the clients get wrong and the line would
soon get to be huge.
...
| WWWWanderer v3.0 by Matthew Gray <mkgray@mit.edu>
| unknown/0.0
| Texture for the X Window System/1.1 libwww/2.12 modified
This table I have to have in all my CGI programs is going to
get huge. I suppose I could have a "I don't know your client,
please click here if you were able to do the following..." but
its getting silly. I've been trying to explain to other staff
in my department that although our application runs on any
platform which has a WWW browser, actually you can't run it
with the browser you just downloaded cause it has a bug - and
not that one as it doesn't authenticate. How are non-WWW
experts to know the difference between browsers?
| Sure, what's so bad about that? Make that table available somewhere
| prominent so others can use it...
What we ought to start then is a definative compliance list! Hopefully
when clients as popular as "WinMosaic" see they are fairly low in
the table they will be motivated to improve!
| 3. We need a way for users to know which client to choose to
| depending on independent conformance testing.
Which does imply that we have standards that everyone adheres too - the
Location/URI http header stuff is all to common. Is this a function
of the new W3O? Sounds like an ideal organisation to test
compliancy and make sure all the standards are fixed. After all, a
browser that says "Tested complient to HTML4, HTTP3 by W3O" is going
to have a market advantage over ones that don't.
Mark
Mark J Cox ---------------------- <URL:http://www.eia.brad.ac.uk/mark.html>
Industrial Technology, Bradford University, UK +44 1274 384070/fax 391333