What's for lunch today?
Everywhere I've worked, a common problem has been that of deciding where to eat lunch. After a while, everyone has tried all possible restaurants and the process of choosing one becomes tedious. "I don't know where I want to eat, where do you want to eat?". Quite a few years ago, I devised the perfect solution to this but only in the last week has that solution been implemented and given a real-world trial. My solution? The Food Wheel. The Food Wheel is a wheel upon which can be placed the names of all the nearby restaurants. Someone spins the wheel. It comes to rest on a restaurant and that's where you eat lunch. Spinning the Food Wheel takes just a fraction of the time taken by the typical "where do eat" discussion.
But until a week ago, my idea was just vapor-ware. Just an idea that got repeated until someone would say, "so, build it already". Last week, Susan took the initiative and turned the Food Wheel into reality. She bought a small lazy-susan (which spawned a few jokes of its own), and used the Dymo labeler, which normally generates little sticky labels for our servers, to put the names of a couple of dozen nearby restaurants on the wheel. And, I'm pleased to report, it actually works. It has been used several times and came up with an acceptable solution every time.
In a strange coincidence, Freshmeat carried an annoucement for a Java applet called "Wheel 'O Yum" this Wednesday which is based on the same idea. It's probably not as much fun as having a physical wheel to spin but it's still pretty cool.
The demise of Red Hat Linux
We've run Red Hat Linux on our servers since version 5. Now we're contemplating what distro to use in the future since Red Hat has announced the end of Red Hat Linux. They will continue to make their high-end Enterprise Linux but it'll cost $1500 - $2500 per server per year. And they're using security updates as a sort of legal loophole that lets them get around the GPL and prevents you from buying one copy of the software and installing it on multiple servers. Supposedly, you can still download the source but if I wanted to compile a whole GNU/Linux system from source, I wouldn't be using a distro like Red Hat in the first place.
I installed Fedora on a test box to evaluate it but it's clearly not intended for serious use in a production environment. It's intended as a bleeding-edge distro for "enthusiasts and developers" only. From what I can tell, it's just a new name for Red Hat's beta distro, Rawhide. Even if someone wanted to use it on a server, it would be impossible - each release is only good for about three months before it's replaced by a new one and support for the old one is dropped. We don't even reboot our servers every three months and I sure don't want to be reinstalling the OS every three months! And then there's the issue of Red Hat trying to trademark the Fedora name and essentially steal it from an existing project.
The only good news on the Red hat front lately is Fedora Legacy, a project that will attempt to offer security patches and support for Red Hat Linux 9 after Red Hat abandons it in April. Looks like we'll stick with that until we can find a new distro. And if some new distro springs up to take it's place, I think they're going to be a lot of takers. I know a lot of folks in the same situation we're in who relied on Red Hat Linux. It's really hard to believe Red Hat did something this stupid. They've effectively found a way to do what none of their competitors had been able to - reduce Red Hat's market share drastically...