[CS-FSLUG] Theory Exercise: Stable Desktop

[veritosproject at gmail.com]

On Wed, Dec 31, 2008 at 9:41 AM, Ed Hurst <ehurst at soulkiln.org> wrote:
> We have developers willing to work under this restriction. With what you
> know about various window managers and desktop environments, what would you
> recommend our team of developers work on? Among the options would be forking
> a current project at some "really good" release point and keeping it there
> (say, KDE 3.1.5), resurrecting some long forgotten favorite (CDE?), or
> adapting some ideas and building a whole new project (a la Xfce).

Building our own, for a very long while (until we got a usable feature
base), would fly as fast as rolling release. Reinventing the wheel
means that you're going anywhere for a while--this is what we saw when
KDE 4.0 came out. The initial release was hailed as garbage, and whole
application groups (kdepim, kdevelop) were missing. It won't be until
4.2 or 4.3 until even the KDE people call it usable.

Having a stable, secure system on (just an example) KDE 3.1.5 would
require backporting five years of work, according to Wikipedia. During
that time, pretty much everything has had patches applied, and
filtering the wheat (security, bugfixes) from the chaff (ooh! shiny!)
would be monumental. On Subversion too, mind you, which compounds the
problem.

I would personally go with an early Xfce4 release---4.4 is simply a
small Gnome. It's stable enough that it functions, small enough that
we can manage it, and (being GTK+) still works with applications well
enough.