If anybody needs a concise example of good coding style, I'd suggest the new epoll code (eventpoll.c) in 2.6... everything explained, especially the "gotcha"s. Very nice work, Mr. Libenzi.
For an example of really *shitty* code, I'd look at some of the drivers. Example: drivers/net/pcnet32.c. All the chip bits and registers are hard-coded w/no documentation. You can't read the code w/out flipping through the technical specs at the same time. Extremely stupid.
What is my incentive to become an Engineer and build things if I can earn the same amount of money being a fry cook.
Honestly? I think I'd rather be an engineer than a fry cook, even at minimum wage. You could be strictly 8-5 (your boss can't make you come in on a weekend when you could just get a job at McDonald's instead) and in theory the pressure would be less.
Don't tell my employer that, though.
What doesn't make sense is going into deep debt in college for a minimum wage job. In our little bizzaro world, most engineers would get their degrees from a community college. (.. if they don't already)
what about hermaphrodites, or other sex organ abnormalities? Where do they fit?
I don't have any local hermaphrodite friends to consult, but it seems fairly obvious that, for not fitting in, they are ostracized accordingly. Would you date one?
So which is better, to promote better female athletes or to promote the visibility of female athletes? I don't know
Perhaps promote a sport where females compete well with males. Like long-distance swimming or somesuch. Possibly boxing but you'd definitely have to mess with the weight classes.
You say that women are generally inherently physically weaker
Well you could also claim women seem to be generally more flexible (looser ligaments and such), and it would be just as "generally" true (meaning only: > 50 percent of the time)
In fact, generally speaking, women with visible muscle are viewed with distaste.
Disagree. Most women (and men) don't go to the gym to be stronger, they do it to look more attractive. What you *will* see is women emphasizing tone over bulk (and vice versa for men). They're different kinds of strength. They're both visible muscle.
I prefer the sink, there are faucets to hold onto.
Oh hell no. If it's a garbage-disposal type arrangement I might see your point. With most sinks you wind up bailing buckets of barf to the toilet. Sometimes in front of cute girls who want to know what's in the bucket(s).
If there's no toilet nearby when the need hits, try a trashbag instead.
:-( You're correct, wrt free alternatives. I meant other proprietary x86 "Unices".
The decision was made years before I got there; unfortunately I can't remember the exact year. It was certainly after the Xenix days. I think the other serious contender might have been
Coherent. Whatever it was, it was lacking in comparison.
I worked on SCO OpenServer at a place that needed a Unix to run their product, and customer requirement was that it ran on PCs. Amazingly enough, in the late 80's/early 90's, SCO was your best bet.
If (s)he's running on a NAT'd network, that shouldn't matter. You'd more likely get in "trouble" when some hotshot starts analyzing the traffic flowing through the corporate/government firewall.
.. and by that last paragraph I did not mean to denigrate or trivialize the hard work that you, ripperda, carmack, and the gang put into utah-glx. Thx.
Nathan (Hand?): I remember the work you did on utah-glx, and am grateful for it. My personal solution to this mess was simply to not buy an Nvidia card.
If you had noted how Nvidia's binary-only module had effectively killed the utah-glx nvidia driver (interesting take on that BTW) a couple of posts up, I think a lot more people would be agreeing with your position.
I've gotta say I'm still sitting on the fence, though. *If* there is still an itch for high-quality, open-source nvidia 3d drivers out there, it will eventually be scratched. All things in their time. Yadda yadda.
IANAaerospace engineer, but I'd guess it's the mostly the aerodynamic profile -> drag; the Harrier is many things but sleek it is not.
I would have also guessed the way the exhaust is vectored even in jet mode might have something to do with it.. certainly that can't be efficient? although if the thrust figures you quote is what's coming out of the nozzle, rather than the raw engine output, your guess is as good as mine.
Lastly, you probably want to be judging based on empty weight, not max weight.
FWIW I read that the Harrier II has approx. 20-22K lbs of thrust (depending on installed engine) and will do Mach 0.98. I understand the F-16 only hits Mach 2. So that's 1/2 as fast, not 1/3.
the vehicle never tries to crank itself the other way.
Yes, but wouldn't the friction between the spinning rotor and the rest of the aircraft still tend to spin the thing? Or is that friction negligible enough not to care?
what good is a counterargument that is just as obviously wrong!?
I mistakenly assumed you were concentrating on "operating systems" innovation. In the general sense you're right, of course. There are plenty of examples of free-software innovation. I would personally cite gnutella, ogg*, slashdot(!), and Mozilla's bayesian spam-filtering as innovative. Freenet has potential. I don't know the exact lineage of BSD's socket API and implementation but it certainly directed the course of history.
And yes, that last one's an example of free-software "operating systems" innovation, but if you'll allow me some wiggle room here, you'll notice I only said we're working on UNIX clones, not that we weren't innovating at the operating system level.. although sockets did happen a long time ago:-)
But the community from which open source software springs is the only place you can get real innovation in the IT Industry. Companies won't try new ideas.
Plan9. Inferno. NeXTStep. Arguably the original Macintosh and the Xerox PARC work they extended.
Meanwhile, in the open-source world, we're fundamentally tweaking UNIX clones. Totally kick-ass, love-'em, best-of-breed unix clones, but still. You might cite Hurd but, comparatively speaking, it's a toy.
None of the commercial UNIX vendors has done a whole lot to advance the state of the operating system from where it was 30 years ago.
Isn't Apple a commercial UNIX vendor now?
SCO in particular, hasn't noticably changed their product since I was saddled with 286 Xenix a decade and a half ago
If your computer is growing slower as your resources increase, it's probably an algorithmic problem. Example: trying to fsck, on a non-journaling filesystem, on disks that grow exponentially in capacity over time. Solution: use a journaling filesystem.
Similarly, you shouldn't need to directly "initialize" all that memory unless it's part of an integrity check (like many BIOS parts do). All that zeroed-out memory would be clobbered anyway as you demand-page programs in from disk.
Some kernels might want to setup page structures or something similar for all known memory at boot time. I figure it ought to be possible to do that in a lazy or delayed fashion instead.
if only Novell had pushed for Linux rather than UNIX in the 90's...
It would have taken a psychic, not just foresight. Novell decided to acquire USL in February '93. Linux 1.0 wasn't released until March '94, and it was (comparatively speaking!) a toy.
If even Novell had known, I don't see how they could have helped Linux in the timeframe they were looking for. Like Mozilla, you knew it would eventually kick ass, but the schedule has a mind of its own.
.. don't know why I'm arguing with an AC (my bad), but alpha-beta pruning has absolutely jack to do with the position evaluation code that constitutes anything resembling a program's "strategy". Alpha-beta is just an algorithm to optimize tree searches.
To better answer the grandparent, position evaluation code varies wildly across programs. Some programs sacrifice evaluation quality for speed and try to make up for it with a deeper search. This would be more akin to the "catch-up" chess you refer to.
Slashdot Mirror
For an example of really *shitty* code, I'd look at some of the drivers. Example: drivers/net/pcnet32.c. All the chip bits and registers are hard-coded w/no documentation. You can't read the code w/out flipping through the technical specs at the same time. Extremely stupid.
Honestly? I think I'd rather be an engineer than a fry cook, even at minimum wage. You could be strictly 8-5 (your boss can't make you come in on a weekend when you could just get a job at McDonald's instead) and in theory the pressure would be less.
Don't tell my employer that, though.
What doesn't make sense is going into deep debt in college for a minimum wage job. In our little bizzaro world, most engineers would get their degrees from a community college. (.. if they don't already)
I don't have any local hermaphrodite friends to consult, but it seems fairly obvious that, for not fitting in, they are ostracized accordingly. Would you date one?
So which is better, to promote better female athletes or to promote the visibility of female athletes? I don't know
Perhaps promote a sport where females compete well with males. Like long-distance swimming or somesuch. Possibly boxing but you'd definitely have to mess with the weight classes.
You say that women are generally inherently physically weaker
Well you could also claim women seem to be generally more flexible (looser ligaments and such), and it would be just as "generally" true (meaning only: > 50 percent of the time)
In fact, generally speaking, women with visible muscle are viewed with distaste.
Disagree. Most women (and men) don't go to the gym to be stronger, they do it to look more attractive. What you *will* see is women emphasizing tone over bulk (and vice versa for men). They're different kinds of strength. They're both visible muscle.
Oh hell no. If it's a garbage-disposal type arrangement I might see your point. With most sinks you wind up bailing buckets of barf to the toilet. Sometimes in front of cute girls who want to know what's in the bucket(s).
If there's no toilet nearby when the need hits, try a trashbag instead.
I call BS. You can make XFree as proprietary as you want (because its license allows it), but that doesn't mean it's not Free Software.
The decision was made years before I got there; unfortunately I can't remember the exact year. It was certainly after the Xenix days. I think the other serious contender might have been Coherent. Whatever it was, it was lacking in comparison.
I worked on SCO OpenServer at a place that needed a Unix to run their product, and customer requirement was that it ran on PCs. Amazingly enough, in the late 80's/early 90's, SCO was your best bet.
The joke used to be that EMC stood for "Excessive Margins Corporation".
I'm assuming they'll have to jack the price just to keep their reputation intact.
We do? I apologise unreservedly. I'll try harder.
You're already screwing it up. You need to try hARRRRRder.
If (s)he's running on a NAT'd network, that shouldn't matter. You'd more likely get in "trouble" when some hotshot starts analyzing the traffic flowing through the corporate/government firewall.
That's an interesting guess, but redhat doesn't think that big yet (because they don't have enough money). How many companies have 50 mil to blow?
Commentators talking around (and insulting) each other is classic slashdot.
Or vice versa..
"damnit grandma, I can't talk right now, it's interfering with a *very* important download.."
.. and by that last paragraph I did not mean to denigrate or trivialize the hard work that you, ripperda, carmack, and the gang put into utah-glx. Thx.
If you had noted how Nvidia's binary-only module had effectively killed the utah-glx nvidia driver (interesting take on that BTW) a couple of posts up, I think a lot more people would be agreeing with your position.
I've gotta say I'm still sitting on the fence, though. *If* there is still an itch for high-quality, open-source nvidia 3d drivers out there, it will eventually be scratched. All things in their time. Yadda yadda.
I would have also guessed the way the exhaust is vectored even in jet mode might have something to do with it.. certainly that can't be efficient? although if the thrust figures you quote is what's coming out of the nozzle, rather than the raw engine output, your guess is as good as mine.
Lastly, you probably want to be judging based on empty weight, not max weight.
FWIW I read that the Harrier II has approx. 20-22K lbs of thrust (depending on installed engine) and will do Mach 0.98. I understand the F-16 only hits Mach 2. So that's 1/2 as fast, not 1/3.
Moderated 'interesting'? More like 'funny' -- FCC != FAA.
Yes, but wouldn't the friction between the spinning rotor and the rest of the aircraft still tend to spin the thing? Or is that friction negligible enough not to care?
No :-) (I'd give you NeXTStep, though)
I mistakenly assumed you were concentrating on "operating systems" innovation. In the general sense you're right, of course. There are plenty of examples of free-software innovation. I would personally cite gnutella, ogg*, slashdot(!), and Mozilla's bayesian spam-filtering as innovative. Freenet has potential. I don't know the exact lineage of BSD's socket API and implementation but it certainly directed the course of history.
And yes, that last one's an example of free-software "operating systems" innovation, but if you'll allow me some wiggle room here, you'll notice I only said we're working on UNIX clones, not that we weren't innovating at the operating system level .. although sockets did happen a long time ago :-)
Um... Linux is only producing toys?
That's *definitely* not what I said :-)
Plan9. Inferno. NeXTStep. Arguably the original Macintosh and the Xerox PARC work they extended.
Meanwhile, in the open-source world, we're fundamentally tweaking UNIX clones. Totally kick-ass, love-'em, best-of-breed unix clones, but still. You might cite Hurd but, comparatively speaking, it's a toy.
None of the commercial UNIX vendors has done a whole lot to advance the state of the operating system from where it was 30 years ago.
Isn't Apple a commercial UNIX vendor now?
SCO in particular, hasn't noticably changed their product since I was saddled with 286 Xenix a decade and a half ago
Sadly, that's probably true..
Similarly, you shouldn't need to directly "initialize" all that memory unless it's part of an integrity check (like many BIOS parts do). All that zeroed-out memory would be clobbered anyway as you demand-page programs in from disk.
Some kernels might want to setup page structures or something similar for all known memory at boot time. I figure it ought to be possible to do that in a lazy or delayed fashion instead.
It would have taken a psychic, not just foresight. Novell decided to acquire USL in February '93. Linux 1.0 wasn't released until March '94, and it was (comparatively speaking!) a toy.
If even Novell had known, I don't see how they could have helped Linux in the timeframe they were looking for. Like Mozilla, you knew it would eventually kick ass, but the schedule has a mind of its own.
To better answer the grandparent, position evaluation code varies wildly across programs. Some programs sacrifice evaluation quality for speed and try to make up for it with a deeper search. This would be more akin to the "catch-up" chess you refer to.
You're probably right; but that's no excuse for never learning to grow a thick skin.