Username or EmailPassword
I think that the issue of how well Gnome runs on old hardware is not useful if the goal is to increase Linux adoption for the desktop.
The advantage of running a Linux desktop on old hardware is adopted from the arguments for running Linux on servers, where the advantage was substituting Linux for other software to do the same work. The fact is, the difference in what a server is asked to do today compared to 5 years ago is far smaller than what a desktop client is asked to do today compared to 5 years ago.
Today someone running a desktop system (and like it or not, this includes every Joe Sixpack who gets his computer from Best Buy) has a far greater expectation as to what that desktop will do for them compared to 5 years ago. Previously, the point was made that used hardware can be bought for $100. This makes no economic sense if one can get hardware that is more than four times as powerful for less than four times that price. Trying to make the argument that one should adopt Linux for the desktop because it will run on old hardware is probably not useful in a market where people are buying new computers rather than trying to fix their old ones.
This is completely redundant. The Joe Sixpacks getting their machines from Best Buy are very unlikely to try Linux anyway, at least for many years. And the consumer market is only a small portion of the computing world.
By having horribly memory-munching desktops and apps like OpenOffice.org and Gnome, the Linux community is locking itself out of a HUGE market: the millions upon millions of machines running NT4 and Win98 in businesses.
Everyone bleats about the 'Microsoft treadmill' and companies aren't pleased with having to buy new hardware for Windows upgrades. And yet these boxes with 64 and 128 MB RAM (if you're lucky) crawl with grossly inneficient software like OOo, Gnome and even Firefox (huge memory leaks). Linux should be there, snapping up a lucrative market, giving businesses a great incentive to switch.
And no, don't waffle on about Fluxbox and Dillo. Large companies aren't interested, and it's a ridiculous solution anyway: to get reasonable speed, you have to ditch the integrated desktop and choose fiddly, feature-lacking programs?!
Again, this'll be ignored, as will the hundreds of posts I see on Linux newbie forums questioning the sheer weight and slowness. Linux's desktop market share will continue to be negligible and everyone will keep pointing at Microsoft rather than real, severe issues that people actually care about (the Gnomers should spend some time with Windows convertees).
It's a lost cause, methinks. Look to Syllable, Haiku and co. to see elegant, efficient coding where the same feature as in Gnome or OOo is implemented 20 times lighter and faster.