Slashdot Mirror
Inside the OpenSolaris Source Code
An anonymous reader writes "Ten million lines of code and not a single profanity? Is that really possible? Apparently, yes, says OpenSolaris community manager Jim Grisanzio. He said even before Sun filtered the code, it was relatively free of profanity. 'They went through the code for a great many things,' he said, 'and I'm sure they cleaned a word or two. Or three.' But a careful look through the code will reveal some programmers' frustration." From the article: "The most embarassing comment came from a developer of the GRUB project who went only by the name of 'Gord'. 'This function is truly horrid,' he wrote. 'We try opening the device, then severely abuse the GEOMETRY->flags field to pass a file descriptor to biosdisk. Thank God nobody's looking at this comment, or my reputation would be ruined.'"
I believe that has nothing to do with the Solaris code.
yep, no profanity at all
Grub is an official GNU project and thus, GPL. Gords comment was intended to be humourous. I'm not surprised the ZDNet hack missed it though. After all, understanding what GRUB is might require that they're are familiar with their subject, and that's just too much to ask these days..
I'm not going to say whether Linux or Solaris is a better OS. But it seems like the Linux code might be a bit more entertaining to read.
I hear Gord dropped out of the video game store market, and is teaching english in Korea.
shit!
Open Source Java Web Forum with LDAP authentication
Good job guys...
He tried to kill me with a forklift!
Most likely this refers to the system tick. On linux this was 100Hz for most architectures in 2.4, although with 2.6 most architectures have moved to 1KHz.
And this coming from the people who gave the world the HME ("Happy Meal Ethernet") network devices?
I guess you're not being too serious.
That's almost certainly the clock interrupt, which, at the time, was generally 100 Hz on Linux, and is now generally 1 kHz. In fact, the thing that's likely to seem quaint before too long is having a constant value, not expecting the value to be less than 10 kHz. This clock is related to system speed, in that it's basically the rate at which housekeeping tasks get done, and it's enough slower than the processor speed that a useful amount of work gets done between ticks, and fast enough that the delay isn't too noticeable when you have to wait for it to tick.
...the Gord in question is almost certainly Gordon Matzigkeit. Make of this what you will.
That would be SunOS, the predecessor to Solaris. Some of the original BSD authors wrote SunOS (and were among the founders of Sun?), so it is not a ripoff.
PPro was a complete architecture redesign. After that it's all been incremental (with the exception of the now-abandoned Netburst architecture). So there's a lot of accumulated changes, but the basic structure and execution approach remain based in the P6. The P6 architecture has proven remarkably robust, surviving the addition of four vector instruction sets, a decade, and an order of magnitude clock speed increase.
Precision is the number of significant figures in a result. Accuracy is how close it comes to a norm (usually as a multiple of the precision). 23.000+/-.001 hours is a very precise but horribly inaccurate measure of the length of a day on Earth--it's off by 1000 standard deviations. By contrast, 23.5+/-1 is far less precise, but more accurate.
So a rock-solid 100kHz clock is less precise (broader timeslices) but more accurate than a drifty 1MHz clock.