JTK wrote:
>
> I don't believe either of those is right, but I'll defer to others more
> intimately involved with the process than I. It is my understanding
> that:
>
> 1. Skinnability was always a top priority. Otherwise, how do you think
> Netscape/AOL was going to get people to work on the project for free,
> yet still 'private-label' it?
> 2. The decision to interpret the GUI was made long after that design
> decision was made.
>
> Are you privy to knowledge that this isn't the case?
Yep. Mike Pinkerton said at an IRC chat I attended back in the early
days of mozillaZine that the Mac port (and this was a chat about the Mac
port) would not exist had not the decision been made to go with a UI
that could be designed cross-platform with one codebase. Netscape at
the time didn't know how much AOL was going to fund them (if indeed at
all), and even now they haven't been given resources to willy-nilly hire
someone for every platform Mozilla is supported on (if it was even the
case that one person could maintain a port to a platform that was
significantly different).
Now, the argument could be made that perhaps an abstraction layer on top
of each toolkit would be better. But that would probably be more work
(especially since it hits a consistently moving target, when underlying
APIs change) than actually writing ones own controls. In any case, the
decision was made because Netscape, when the decision was made, couldn't
afford to maintain full GUIs for each platform (and they probably still
can't today; AOL will only pay for so much). Skinnability wasn't even
(easily) possible until several months after XUL first appeared (it was
one of the big beta2 tasks for N6).
