First I wanted to point out two things you seem to be ignoring -
1. Celeste is no longer in the image. I posted a script for it's removal
a while ago, and it was subsequently patched and rolled into the update
stream for many of the reasons you mention.
2. Adam Spitz has taken over maintainance of Celeste, and has already
integrated a few patches that had been laying around.

I would definitely love to see you and Adam work out how to come up with
one core Celeste that has the best features of both, maintained by one
person. Anyway, now that Celeste is out of the image, it should not be
too hard to go ahead and separate the UI from the model (which would not
be a bad idea anyway), and then maintain two UI based on the same model,
if we find the preference differences are hard to reconcile.

Daniel

Lex Spoon <lex at cc.gatech.edu> wrote:
> 
> I do prefer the UI.  The UI is identical except for how it deals with
> filtering.  Instead of having a button for each type of filter, FC gives
> you a generalized list of filters.  FC's filters may include the usual
> participant and subject filters, plus a new "category" filter to replace
> the top-left pane of classic Celeste.  The custom filters are still
> there, but they've been renamed to "named filters".
> 
> That's it.  This UI is still fast to use for the usual case of one
> category filter, one named filter, and one subject or participant
> filter.  Now, however, you can have multiple filters of the same kind if
> you want.
> 
> I actually envisioned FC letting you combine multiple filters with AND,
> OR, NOT.  In particular, it's lame to resort to a "code" filter just
> because you want to do a participant filter on multiple possible
> participants, e.g. to filter out all email that is addressed to any of
> your personal email addresses.  Sadly I don't have time to work on stuff
> like this.
> 
> Aside from this there are a lot of small goodies that I don't think ever
> got into the main image's Celeste.  From the top of my head:
> 
> 	- It is speedy on large folders
> 
> 	- It interacts correctly with image snapshots
> 
> 	- It lets you open multiple windows at once on the same mail database
> 
> 	- It does some save's more quickly, due to putting Daniel's (I think?)
> logging idea straight into the index file
> 
> 
> Then again, I'm sure a patch or two has gone into main Celeste but not
> into FilteringCeleste.
> 
> 
> I never intended for there to be a fork.  The thing is that no
> individual is able to make major changes to the "main" Celeste.  There's
> no designated maintainer, Squeak Central doesn't maintain it, and the
> harvesters are too conservative.  So FilteringCeleste has become my own
> release of Celeste, much like Snail on Hot Asphault (SHA) was my own
> release of the Unix VM back when Ian Piumarta was busy doing other
> things.  I happen to think I've done a great job on it :), but it's up
> to the community what to do next.  I'll personally keep using FC until
> there's a better alternative.
> 
> IMHO it is a waste of time to do non-trivial work on "main" Celeste.  FC
> is superior in many ways, and even the UI can be switched if it is a
> real roadblock for people.  Celeste only gets modified right now by
> going through the harvesters, and this is not a process that allows
> major changes.  If some maintainer does step up and start working on the
> main Celeste, I still think the best thing would be to start by
> integrating the improvements FC makes.  However, that effort will
> probably be more than the effort of adding a classic UI to FC.  FC is
> simply much more about non-UI things than UI things, despite its name.
> 
> 
> I never thought I needed to say it explicitly, but I'd certainly be very
> happy for FC to be considered "the" maintained version of Celeste.  I
> would integrate patches -- or negotiate with people until I would accept
> the patches -- and I'd make an occasional release.  Celeste hackers
> would love this, I think; even with my limited time, it should be much
> faster to get things posted through me than through the harvesting
> process.  Further, I can allow much bigger changes to go through than
> will make it through into the main image.
> 
> Let me put this another way.  Having a single maintainer is a good for
> an open source project.  A maintainer can move things along more
> quickly, and a maintainer can keep the program coherent.  I don't insist
> on the maintainer being me, but I'm a natural choice and a good choice. 
> Natural, because I got Celeste into Squeak to begin with and because
> I've been maintaining it ever since then.  Good, because I know the code
> extremely well, I have a good design sense, and I think I'd be pleasant
> to work with.  While my pickiness (to be kind) about design would irk
> some people, I'm quite willing to go through iterations with people.
> 
> On the other hand, I have limited time.  If someone more eager wants to
> jump in and take over FilteringCeleste, that would be great.  I'd like,
> however, to at least see some patches from them before handing over the
> baton, and I haven't seen a single patch as yet.
> 
> 
> In sum, both general users and more devoted hackers should be much
> happier with Filtering Celeste than the Celeste in the main image. 
> Switch over, already.  :)  Squeak Map makes it very easy.
> 
> 
> Lex