RE: [cobalt-developers] Problems with the Cobalt GUI and Linux

[Ian McCall <ian@xxxxxxxxxxxx>]

"Robert Walter" <rwalter@xxxxxxxxx> wrote:
>...If alternate browser developers can't keep the
>basic functionality of the HTML standard intact, should every other
>developer in the world dance to their drum?

No. And that's the point. It's IE that deviates from the standard. To test in
the most standards-compliant browser at the moment, use Mozilla (or any
derivative of the 1.0-codebase). Then tweak for IE.

Amongst other things, I also do web development. If you start out with a
non-standard page and try to force it into the standard (the "write for IE and
tweak" approach), it is always harder than writing to the standard and then
applying little tweaks to get it to work in the non-standard ones.

Actually, IE is pretty good at the standards these days so I'm surprised this is
a problem. They do have a few hassles (decent PNG support would be nice), but
then so do the other side (COLGROUP not handled very well in Mozilla).

To be honest, I'm quite shocked and depressed at the line coming out that says
Cobalt test only with the mass-market browser. My opinion - they should test
against the defined standards, and then adapt to the mass market -if it proves
necessary-. Mostly, it won't.


>Get the guys that write the browser software on the same page!

Well, effectively there are about five pages.
 - The standard (best exemplified by Mozilla and its derivatives)
 - IE (little variance these days)
 - Netscape 4.7 (horribly broken, may a curse be laid upon its CSS support for
all time...)
 - Konquerer
 - IE for the Mac (small differences, but they're there).

To be completist, there's also Opera (cross platform) and Omniweb (Mac OSX).

Just to restate - my approach is to always develop to standards using Mozilla to
view as I write, then to test with IE to ensure compatibility. I'm surprised to
hear a company, particularly a Linux-based company, argue that writing to
standards is not important.


Cheers,
Ian