OpenBSD's Alpha Support In Trouble
Nimrangul writes "Hours ago Theo de Raadt put out a call for an Alpha CS20, because as of last night OpenBSD no longer has one. The CS20 that died was a build machine and without it further support for the Alpha platform would be nearly impossible. If you have a C320 or other 1U Alpha machine that you would be willing to donate to the project, please respond to the discussion on the misc mailing list."
Given the amount of equipment in Theo's server room and given the importance of this equipment to the project, why not construct a thermal shutdown device? How about a machine with a number of temperature probes around various points in the room, and when they all agree that the temperature is hot, they initiate shutdown+power-off procedures on the machines in the room? Now, I realize that some of the machines in the rack are older and may not have self-power-off abilities but it seems likely that enough of them could power down to make a difference.
The lack of actual compiling on your fringe hardware is why the support for it is so bad.
While it is true that it can be compiled faster on other hardware, that doesn't mean that the machine itself can compile it's own copy of the operating system.
If my machine cannot compile it's an operating system supposedly designed for it, there is a problem with the code and most likely how it works.
Cross compiling can be handy for speedy development, but not quality development. That's where the actual hardware comes in.
I find it amusing that you'd suggest nobody uses NetBSD at a time when the front page of slashdot carries a link to a quaterly NetBSD report mentioning seven new developers, seven google "Summer of Code" projects and a number of donations from both individuals and corporations.
Just because we don't make such a song and dance over it doesn't mean we don't exist.