Unfortunately, open source tends to be better at Battlestar and less bad at Gunsmoke.
Volunteers just aren't going to get excited about the task of updating a knowledge base every year for the new tax code; it's generally more exciting to do something that's new.
Now, you could say that Linux is more like Gunsmoke, in fact, it's more like Gunsmoke than Windows or Mac OS X is -- and because an operating system is something that you depend on day in and day out, an OS ~should~ be like Gunsmoke. Linux, however, is an unusual case. Today Linux development is dominated by a few big corporations that either sell Linux commercially (Red Hat) or that decided that, by contributing to Linux, they could have a better Unix than they could develop in house (IBM)
The Gunsmoke trend will continue as long as customers keep changing their minds about what the heck they think is important, or because the software needs more power to satisfy how the users actually work with it.
That certainly drove the early releases of MS Word, Quicken, Photoshop and others. Eventually, there's "enough" to work with (feature wise) and the milking really begins because the need to upgrade is more perceived than really needed...The last few releases of Word and Quicken certainly demonstrate this well.
Eclipse platform is a good example of this problem. Ever since the resource perspective stopped being the default, things have been headed down an uncertain path.
Why can't there be a true core, that has some chance of being 'finished' at some point?
So should scientific and technological progress in general be like Battlestar? I would much rather it not have a known ending. If I wanted that, I'd choose religious dogma.
I think the analog in science are people who just keep doing the same old studies slightly better (ideally), publishing papers like "This year we've managed to beat down the error on parameter X to 1/3 of what was previously possible." even though it's not clear what the benefit of knowing X better is apart from assuring a continuing train of grants supporting them.
Volunteers just aren't going to get excited about the task of updating a knowledge base every year for the new tax code; it's generally more exciting to do something that's new.
Now, you could say that Linux is more like Gunsmoke, in fact, it's more like Gunsmoke than Windows or Mac OS X is -- and because an operating system is something that you depend on day in and day out, an OS ~should~ be like Gunsmoke. Linux, however, is an unusual case. Today Linux development is dominated by a few big corporations that either sell Linux commercially (Red Hat) or that decided that, by contributing to Linux, they could have a better Unix than they could develop in house (IBM)