Whether Moz should split into Firebird and Thunderbird or stay a suite is still, apparently, up for debate: "
>> Brian, I did the math. 18.7 M (Firebird/Thunderbird) to 15M Mozilla
>> App/Bloat Suite. These are compressed not extracted footprint numbers :-)
>> Now we have everything as a 'stand alone' distro which is *increasing'
>> the bloat not reducing it.
>
>That's just because it is still a work in progress. Also in the future a
>shared GRE will reduce this further.
>There could easily be a release that includes both Firebird and
>thunderbird as standalone apps, but sharing the same GRE, which would be
>smaller than the current 15M"

Boy, there's a great idea. Split the suite, then recompile the suite by sharing innards.

It sounds like there is another project in here somewhere -- specifically that Mozilla's core needs to be rewritten seperately, and then you can lay applications on top of that (read: Thunder & Fire 'birds). I think they'd gain some advantage checking out the Mozilla ActiveX control project. Yes yes yes, I realize that's Windows only, but the point is well taken and could be extended to DCOM.

Make an engine, a DCOM object, that supports a very simple API and make it easy to build around that. A browser would be child's play, the way making a browser with VB6 and the Microsoft Internet Control (or the Moz ActiveX control) is now. Thunderbird would use this plugin for message display... and probably message composition. Let other UI choices belong to the two projects, from menus to sidebars to... That's all nasty overhead you don't need in the core object, which should concentrate on great rendering of standards, speed, and stability. And, afaict, Mozilla's code's setup already favors the break.

So I think there are three projects -- a browser, an email client, and a refactored, common, core rendering engine. What's more, the browser should keep the name Mozilla. There's too much name recognition there -- and name recog that's browser-specific -- to throw away.

Oh well. Random rant.