> It has been commented that I said nothing about Composer and Mail/News.
> There are primary reason for this is that I know next to nothing about
> these two areas, and they are not really mentioned in the original 1.0
> doc. Still, I should have cross-posted the message. So, folks, come and
> join the discussion.
I think Mail should meet the criteria listed as "target" in MailNews
performance tests, or at least come very near (stay, e.g within 20% of
this target numbers)
For functionality of MailNews, I think most of the unresolved 4xp bugs
should be fixed, and POP/IMAP should work correctly (from my experience
with POP, this is already going well).
For News, I think the most important missing point is false numbers of
unread messages for threads, people won't accept this in 1.0 software.
the rest of this is at a quite good level from what I see (of course,
some performance issues left but they should be covered by the tests
mentioned above).
> On that point, though, there's an issue to be raised. Users will expect
> Mozilla 1.0 to be a "function" release - a certain level of functionality
> and stability. Distributors may be looking more at API freezes and
> embedding stuff.
I think users and developers will expect this to be better than NS4.x in
terms of functionality, so I think remaining 4xp bugs should be one of
the main targets after stability, performance, dataloss, and embedding
criteria. We should perhaps set a line for "this many 4xp bugs can still
remain" or better say "4xp issues need good reasons to be pushed after
1.0, and pushing them off must get approved by drivers" or similar.
> Are we looking at the function of the apps built on the Mozilla framework
> for 1.0, or the framework itself? Or both? Where's the line?
As the browser suite is easier to compare with other products, and will
be the first target of use by other people, we should focus our criteria
on this one, I think. The framework should be OK with an API freeze,
IMO. Developers then have APIs they can build upon, and I think the
relatively young framework will still need some time to grow underneath
those APIs. We will have on a growing number of applications built on
that framework, and that is only possible if they have stable APIs to
build upon.
So I'm getting five important target areas for 1.0:
stability, performance, dataloss, embedding/APIs, and 4xp.
if we set up some criteria in those areas, I'm satisfied (others may
think completely different and I'm noone really important)
;-)
Robert Kaiser
contributor for L10n, skins, and RBase
