Subject: RE: Topic: Just Works and 1.0
From: Paul Rohr (paul@abisource.com)
Date: Fri May 04 2001 - 15:54:54 CDT
Bill,
Thanks for your *superb* clarification below. I'm assuming that you share
my sense that 1, 3, and 2d-ii are to be avoided if at all possible, no?
If so, then I'm more than happy to discuss any features which are clearly in
the 2d-i category. In general, I tend to focus my efforts on getting
features to Just Work for a given release. If that means displacing the
entire feature to a later release, then so be it. In many -- perhaps most
-- cases, I can live with that.
However, releasing features which are clearly in category 2d-i is less
scary. As you say, if the first half of the feature really does Just Work,
then most reasonable people will be willing to wait one release to get the
other half in the same state. That's a promise they'll only buy once,
though.
If we're not in 2d-i, then it's very very hard to know what it'll take to
guarantee that the rest of the feature *will* in fact be ready for the next
release, and that just sucks. I don't want to get stuck in that situation.
If we're gonna make any #2 promises, then I want to be very sure we know how
to deliver on them.
Paul
At 10:19 PM 5/3/01 -0700, WJCarpenter wrote:
>This is not enough granularity, because different people have
>different perceptions of what "half-works" means. I doubt that
>"half-works" and "just works" are opposites. Here are three
>definitions that immediately spring to mind. It wouldn't surprise me
>if others on this list were actually thinking something different from
>all of them.
>
>1. "half-works" means it has a bunch of bugs, maybe crashes, maybe
> doing non-intuitive and/or irritating things instead of what
> people really want.
>
>2. "half-works" means "not complete". For example, if we had a
> tables implementation that didn't let you vary the thickness of
> the border lines, that would be a surprise to some users, but some
> percentage of them would say, "I don't care about that" and some
> other perecentage would say, "I care about that, but I guess I can
> live without it for a while longer".
>
>3. "half-works" means "not complete" like above, but the
> incompleteness is more random and unpredicatable. If we didn't
> have variable table borders, that would seem incomplete. If we
> had variable table borders except on the left sides of cells and
> tables, that would be incomplete and kind of incomprehensible no
> matter what technical challenge led to it.
>
>There's are variations of #2 that are the developer versions of the
>given user version. They are:
>
>2d-i: Incomplete, but there is an architecture that makes it look
> like a simple matter of coding to get all the way there as time
> permits.
>
>2d-ii: Incomplete, and the architecture makes it look like adding the
> rest of the feature will require a lot of ripping and
> rewriting.
This archive was generated by hypermail 2b25 : Sat May 26 2001 - 03:51:02 CDT