If you're doing Power and stuff, you might want to go for the PE. If you're doing more of the microelectronics it doesn't matter very much. The PE test is geared much more towards ME/CE/ChemE than it is EE/CmpE.
Most people from here at Georgia Tech who graduate with an EE or CmpE degree don't take the PE test.
Exactly. And unfortunately, this will only exasperate that problem. It encourages "drop 'em in front of the TV and let 'em bake for hours" parenting.
I disagree. I don't think that this will make the problem much worse than it already is. I think that Children's shows have already made the problem (Seasame Street is now Daddy)... I think that parents who tend to neglect their children will continue to leave their children in front of the T.V... this will just give them more of a piece of mind. I think that parents who still spend lots of time with the kids for the most part won't be largely affected by this... but it will help, maybe, when the baby sitter you don't know comes over and wants to watch BrutalDeath III, or when your four-year-old wakes up and slips downstairs to watch whatever is on HBO at 2:00 in the morning.
Anyway, it's a purely optional thing and so I don't think it's so bad.
I think that the biggest problem in most of modern culture is not that we have too much violence on tv but that our parents don't spend enough time with their children. That being said, is the V-Chip so bad? Sixteen-year-olds are able to get their parents to give them the password, or are able to get their own T.V. So their rights aren't infringed. Three-year- olds aren't able to watch blood and gore on TV. Is that so bad? Shows that are designed for Adults but which appeal to kids for the wrong reason (Southpark comes to mind) are at least a little harder for kids to get into. And all of you who are old enough still get to watch it. What's so bad about it again?
Why? Because of competition. We see neato processors from Sun and Motorola/IBM because they need to keep ahead of Intel. We had lots of innovation during the Space Race because we needed to keep ahead of the commies. We have lots of innovation in the military because we want to remain the best. What do these things have in common? They're all competetive. And perhaps the stakes are a little higher when you're talking about the military... if you don't have the best technology, you die.
There has been lots of work for the Matrox G[24]00 cards recently. Matrox has released nearly all (if not totally all) of the specs for the cards (hundreds of pages worth) which makes programming for them even easier than the TNT cards (which release source but not the best in the way of specs). With some of the most recent code for X / GLX / AGP / Gx00 WARP and stuff, we are reaching pairity with Windows -- according to John Carmack, we have even better performance than Windows on Quake2demo2, and have about 90% of the Windows Quake3 demo speed. And we're improving all the time. So go for those G400 boards... they're darn fast and getting faster!
Righto... I knew that. But based on what we've been sending to other planets recently (using, say, the mars probes as the state of the art) we aren't getting all that much bandwith (in addition to high latency) due to all sorts of factors, not the least of which is that you have to spend lots and lots of space on Error Detection & Accomidation. And if we, say, have a megabit between here and Neptune, we would want to save that for people who needed it, not slashdotters using 10,000 ping samples to estimate the current distance from here to the planet.
I'd guess that ping flooding wouldn't work... Fiber is cheap, but inter-planetary bandwith will come at a premium, is my guess. There will probably be all sorts of things to make sure only packets from authorized places will make it between planets, and also bandwith constraints.
As it happens, FTP is already for the University of Mars. Just not telnet.
Re:Technical Question: Scrollbar placement.
on
The Future of KDE
·
· Score: 1
In theory, this is themable in GTK 1.2. You can also (again, in theory) put your menubars in a common area, a la Macintosh.
Gnome isn't at a 2.0 level yet, no. Neither is KDE... that's why both Gnome 2.0 and KDE 2.0 are in development. And Gnome will be 1.2 before it is Gnome 2.0 (1.0 with most bugs worked out). Do you remember using Linux 1.0.x? There were quite a few quirky things in there, too. Gnome is still sort of unstable, yes. There are still some big issues (esp. when trying to port to other OS's... I still can't get corba-gmc to compile under solaris). But all in all, I must say that Gnome has a bunch of nice features. I can drag from gmc into xmms or gqmpeg (on my linux box, where gnorba works).
Nevertheless, I have some problems with Gnome. It's not stable. Many (most?) of the programs lack standards.
There are two big problems I have with KDE. 1) I think it's ugly -- but that's just me. 2) KDE is tied into Qt and C++.
So, I would say that my problems with Gnome are that Gnome has lots of bugs. But my problem with KDE is an architectural one. I can't write Ada programs for KDE. I can't write graphical perl programs for KDE. I can't write C applications for KDE. I can't write python applications for KDE. Only C++. C++ may be a good language, but sometimes it's not the right tool for the job. And that is perhaps that is the most important thing in programming, computing, and life in general: always use the right tool for the job.
Now, don't get me wrong, I think that it's great that the KDE developers have done such a good job. I hope they continue their work as long as they get joy out of it. But when I develop my graphical applications, I'll do so for Gnome, because Gnome is more flexible, and I have more choices.
SuSE has alwaysbeen the most reliable and stable commercial Linux distro,
No, that would go to Slackware. That's why slackware is still in libc5 land (which will change RSN, Slack5 is nearing beta readiness).
It is common to ship Linux with the latest stable kernel, unless there are *real* showstoppers.
I would call the file corruption bug a ``showstopper.'' Perhaps SuSE isn't on the linux kernel-dev mailing list. They should be. SuSE has also shipped a snapshot version of XFree in the past, and instead of telling users to upgrade to a stable version when X had problems, they forwarded bug reports to the XF86 group. That is why XF86 development is no longer open (IE, you can't get to the in-development source trees). Now they are using 3.3.4... which (in the release notice... which you would think SuSE people would read before throwing it in the distribution) has been declared to be an intermediate, non-production-quality release.
Look here for the article from the Alumni Magazine. There's a nice picture, and they discuss some of the cooler uses, like looking at the heart rate for rifle people in the olympics.
My server is a P100 with 64 megs of RAM and two 6.4G hard drives RAID-1'ed. I had (at one time) four people using MATLAB (with fairly good speed) from remote locations (using ssh, of course) I had open 5 or 6 xterms (from my workstation, the server has X libs but no X server) and was compiling the kernel. One other user was reading mail. Yes, I was hitting swap, but not too badly, and everyone was running with fairly good speed -- pine seemed as fast as normal, but my shell had swapped out by the time I exited, and I had to wait a tad...
In any case, Linux does a really great job with scheduling. Processor power is something that is useful for raytracing, rendering, and gaming, but a fairly recent processor (low-end pentium) can handle even several processor-intensive applications (such as MATLAB) amazingly well. The best upgrade for your linux box is more RAM in most cases, I'd say. That's where most of my slowness comes from: swapping, especially in X, when you can see (And hear) some window being sucked off the disk...
You can get things in windows that add windows explorer capability for some different kinds of files -- including ftp and zip files. But it's not nearly so elegant... you have to use different programs for each type of thing you want to browse (rather than just link in a new library from MC) and my experience has been if you try to do things like open a zip file from an FTP site that things go haywire. I'm not sure that the system is actually implemented with real filehandles, either; it might just pass a temp file into the program. In any case, podfuk is really nifty and elegant. And I don't use windows enough to discuss the system-level details of the winzip or web folders extensions.
I think two+ is a sign that a movement has reached commericial mainstream. So, uh, congrats to....someone.
The congrats goes to us. Every one of us who has written a patch, reported a bug, advocated using linux in the right place. Every one of us who has written up some documentation or helped a friend do an install. Every one of us who answers newsgroup posts or ask slashdot posts.
Many of us have put effort into Linux (and free software in general), and we are seeing people take notice. We have built it, now they are coming. And as corporations start to use our software, they will need their own itch scratched, and Free Software will continue to grow and become stronger and more flexible.
The other day my friend showed me a really nifty piece of software called podfuk. It basically (so far as I can tell) combines vfs libraries from Midnight Commander and the CODA filesystem... so you can cd into tar files or ftp sites, and the actual work is done in userspace. It's a very, very elegant hack, and I think it is exemplary of the quality and flexibility of the software that we have. I don't even want to think of how you could do that on an NT or Mac machine.
So, give yourself a pat on the back. We've come a long way. And we are getting the fame that we crave.
Now, get back to work. We still have lots to do.
Re:Without a SysRq I would have a dead machine tod
on
Changing the Keyboard
·
· Score: 1
So, I have some issues with all sorts of things with the ``Mac Interface,'' but I'll just talk about how nice the sun keyboard is instead.
No. Actually, *Command-C*, -V, -X, etc from the Mac is the proper way to do this. That way you hit the modifier key with your thumb and the key with your middle or pointer finger. Very comfortable. M$ screwed things up when they ripped off the Mac, and used Control instead. Not nearly as nice, given the standard Control position (which I keep...I don't swap with Caps Lock).
Ick, you like having the modifier key right below the key you're trying to hit? Anyway, you really should switch caps lock/control (or, better yet, make both caps lock and control control, and have no caps lock, which is what I do). Control-C (aka break) Control-V (quote-character) Control-X (in Pine) Control-W (werase) and most other control sequences are all so nice with the control under the tab, because your pinky is already there... I loathe having to move my pinky down two keys...
Get a copy of xkeycaps and duplicate some keys today!
I don't like a one-button help key...again, easy to accidently hit and start up an annoying help system. Better to use command-? or something.
Hmmm... but having a big button there is nice... back when I had to use a Sun5PC keyboard, before I got my Sun5Unix, I had a much larger area in which I could thwap my escape key. The way my keyboard layout is, I can hit escape, null, or help, and they all give me escape. I can sort of push my hand in that general area of the keyboard, and it will work... useful if I'm coming from an icky PC or Mac or something. Anyway, I miss the escape much less when it's above the tab then when it's an inch above the tilde. And I like having the backspace right next to the close bracket... pinky has to move less far.
I found some (poor) plans on the 'net for one, but required almost 100 USdollars worth of parts and programming a microcontroller. I'm taking a microcontroller class at school this fall and I'll see if I can re-implement it for (hopefully) cheaper....
Also, a few multiport adapter companies have converters for sun*PC*... you could use that, but they're more expensive ($150+USD)
But what most people don't understand is that *nix is easier on the desktop for most companies, given that the software they need is available (if they need Word, then obviously *nix might not be the best choice). If you drop Windows98 on everyone's desktop, soon they are going to install some nifty utility, and then a screensaver, and then realize that the Screensaver needs DirectFoo5.2 and soon their computer won't boot. So you have to send a tech out to fix it...
The nice thing about *nix is that you build the desktop system, you set up the user with a nice menu on the root window or on a button bar, and as long as that user doesn't know the root password, the user can't mess up the computer. That's why our PC group here has over 8x as many people as our unix group... not becuase there are more PC's (the number is roughly the same, plus the servers are mostly UNIX), but because they are so much harder to maintain.
For the typical home user, or for small businiesses that just want a word processor or spreadsheet, Windows or MacOS is still probably their best choice -- the learning curve for Linux is probably not worth it (yet). But for a company that has to hire IT people and don't need windows-only programs, *nix can lower costs and problems a lot.
Uh, I do check the checksum. If Linus proves to be ineffective as kernel maintainer, then he will be replaced. He has so far, however, proven himself to be fairly good at knowing what goes in and what stays out.
That's one of the things maintainers do: filter out all the crap code. That's why projects have maintainers. That's why we don't have a bazillion trojans in the code.
This further supports theories that SGI will be moving to Linux... People buy SGI workstations because they run Maya and company, and run them well. Linux provides not only an efficient, stable environment to work in, but it will run pretty much the same way under a Mips 10k as it will an x86. That's a huge advantage over Irix.
Now if only they give us some of the nice kernel multithreading that Irix has, and that 4dwm-zooming-filemanager...
But with simple constructs added to languages, having the compiler create parallelized code would be much more trivial. For instance, when you want to do something 100 times in C, you would do
int i,foo[100],baz[100]; /* stuff in here */ for (i = 0; i/* Some stuff in here not dependant on */ /* Previous iterations */ /* We'll add foo and baz */ foo[i] += baz[i]; }
Now, that's silly. You don't want to do it for i=1, then for i=2, then for i=3, you just want to have it done 100 times, for values between 0 and 99. Adding two arrays, for instance, or zeroing them out. So, we need a construct such as
multido(i,0,99) { /* that stuff in here */ foo[i] += baz[i]; }
Now, in most current and traditional program designs, this would still be non-trivial, as the compiler would have to start up two threads, let them process, wait for them to shut down, and so on, but in an architecture where overhead was less (pre-existing threads where you just pass code for them to run, maybe?) things like the above could be extremely beneficial.
Uh, didn't they say 6,000,000 shares? Doesn't that mean 60 million dollars worth? Or is the 6,000,000 just what they are releasing and doesn't accout for other investments?
Slashdot Mirror
Most people from here at Georgia Tech who graduate with an EE or CmpE degree don't take the PE test.
WINDOWS NT SERVER 4.0 IS ENTERPRISE-READY :-)
I disagree. I don't think that this will make the problem much worse than it already is. I think that Children's shows have already made the problem (Seasame Street is now Daddy)... I think that parents who tend to neglect their children will continue to leave their children in front of the T.V... this will just give them more of a piece of mind. I think that parents who still spend lots of time with the kids for the most part won't be largely affected by this... but it will help, maybe, when the baby sitter you don't know comes over and wants to watch BrutalDeath III, or when your four-year-old wakes up and slips downstairs to watch whatever is on HBO at 2:00 in the morning.
Anyway, it's a purely optional thing and so I don't think it's so bad.
I think that the biggest problem in most of modern culture is not that we have too much violence on tv but that our parents don't spend enough time with their children. That being said, is the V-Chip so bad? Sixteen-year-olds are able to get their parents to give them the password, or are able to get their own T.V. So their rights aren't infringed. Three-year- olds aren't able to watch blood and gore on TV. Is that so bad? Shows that are designed for Adults but which appeal to kids for the wrong reason (Southpark comes to mind) are at least a little harder for kids to get into. And all of you who are old enough still get to watch it. What's so bad about it again?
Why? Because of competition. We see neato processors from Sun and Motorola/IBM because they need to keep ahead of Intel. We had lots of innovation during the Space Race because we needed to keep ahead of the commies. We have lots of innovation in the military because we want to remain the best. What do these things have in common? They're all competetive. And perhaps the stakes are a little higher when you're talking about the military... if you don't have the best technology, you die.
There has been lots of work for the Matrox G[24]00 cards recently. Matrox has released nearly all (if not totally all) of the specs for the cards (hundreds of pages worth) which makes programming for them even easier than the TNT cards (which release source but not the best in the way of specs). With some of the most recent code for X / GLX / AGP / Gx00 WARP and stuff, we are reaching pairity with Windows -- according to John Carmack, we have even better performance than Windows on Quake2demo2, and have about 90% of the Windows Quake3 demo speed. And we're improving all the time. So go for those G400 boards... they're darn fast and getting faster!
Righto... I knew that. But based on what we've been sending to other planets recently (using, say, the mars probes as the state of the art) we aren't getting all that much bandwith (in addition to high latency) due to all sorts of factors, not the least of which is that you have to spend lots and lots of space on Error Detection & Accomidation. And if we, say, have a megabit between here and Neptune, we would want to save that for people who needed it, not slashdotters using 10,000 ping samples to estimate the current distance from here to the planet.
As it happens, FTP is already for the University of Mars. Just not telnet.
In theory, this is themable in GTK 1.2. You can also (again, in theory) put your menubars in a common area, a la Macintosh.
Nevertheless, I have some problems with Gnome. It's not stable. Many (most?) of the programs lack standards.
There are two big problems I have with KDE. 1) I think it's ugly -- but that's just me. 2) KDE is tied into Qt and C++.
So, I would say that my problems with Gnome are that Gnome has lots of bugs. But my problem with KDE is an architectural one. I can't write Ada programs for KDE. I can't write graphical perl programs for KDE. I can't write C applications for KDE. I can't write python applications for KDE. Only C++. C++ may be a good language, but sometimes it's not the right tool for the job. And that is perhaps that is the most important thing in programming, computing, and life in general: always use the right tool for the job.
Now, don't get me wrong, I think that it's great that the KDE developers have done such a good job. I hope they continue their work as long as they get joy out of it. But when I develop my graphical applications, I'll do so for Gnome, because Gnome is more flexible, and I have more choices.
No, that would go to Slackware. That's why slackware is still in libc5 land (which will change RSN, Slack5 is nearing beta readiness).
It is common to ship Linux with the latest stable kernel, unless there are *real* showstoppers.
I would call the file corruption bug a ``showstopper.'' Perhaps SuSE isn't on the linux kernel-dev mailing list. They should be. SuSE has also shipped a snapshot version of XFree in the past, and instead of telling users to upgrade to a stable version when X had problems, they forwarded bug reports to the XF86 group. That is why XF86 development is no longer open (IE, you can't get to the in-development source trees). Now they are using 3.3.4... which (in the release notice... which you would think SuSE people would read before throwing it in the distribution) has been declared to be an intermediate, non-production-quality release.
From the labs at Georgia Tech Research Institute.
Look here for the article from the Alumni Magazine. There's a nice picture, and they discuss some of the cooler uses, like looking at the heart rate for rifle people in the olympics.
In any case, Linux does a really great job with scheduling. Processor power is something that is useful for raytracing, rendering, and gaming, but a fairly recent processor (low-end pentium) can handle even several processor-intensive applications (such as MATLAB) amazingly well. The best upgrade for your linux box is more RAM in most cases, I'd say. That's where most of my slowness comes from: swapping, especially in X, when you can see (And hear) some window being sucked off the disk...
You can get things in windows that add windows explorer capability for some different kinds of files -- including ftp and zip files. But it's not nearly so elegant... you have to use different programs for each type of thing you want to browse (rather than just link in a new library from MC) and my experience has been if you try to do things like open a zip file from an FTP site that things go haywire. I'm not sure that the system is actually implemented with real filehandles, either; it might just pass a temp file into the program. In any case, podfuk is really nifty and elegant. And I don't use windows enough to discuss the system-level details of the winzip or web folders extensions.
The congrats goes to us. Every one of us who has written a patch, reported a bug, advocated using linux in the right place. Every one of us who has written up some documentation or helped a friend do an install. Every one of us who answers newsgroup posts or ask slashdot posts.
Many of us have put effort into Linux (and free software in general), and we are seeing people take notice. We have built it, now they are coming. And as corporations start to use our software, they will need their own itch scratched, and Free Software will continue to grow and become stronger and more flexible.
The other day my friend showed me a really nifty piece of software called podfuk. It basically (so far as I can tell) combines vfs libraries from Midnight Commander and the CODA filesystem... so you can cd into tar files or ftp sites, and the actual work is done in userspace. It's a very, very elegant hack, and I think it is exemplary of the quality and flexibility of the software that we have. I don't even want to think of how you could do that on an NT or Mac machine.
So, give yourself a pat on the back. We've come a long way. And we are getting the fame that we crave.
Now, get back to work. We still have lots to do.
No. Actually, *Command-C*, -V, -X, etc from the Mac is the proper way to do this. That way you hit the modifier key with your thumb and the key with your middle or pointer finger. Very comfortable. M$ screwed things up when they ripped off the Mac, and used Control instead. Not nearly as nice, given the standard Control position (which I keep...I don't swap with Caps Lock).
Ick, you like having the modifier key right below the key you're trying to hit? Anyway, you really should switch caps lock/control (or, better yet, make both caps lock and control control, and have no caps lock, which is what I do). Control-C (aka break) Control-V (quote-character) Control-X (in Pine) Control-W (werase) and most other control sequences are all so nice with the control under the tab, because your pinky is already there... I loathe having to move my pinky down two keys...
Get a copy of xkeycaps and duplicate some keys today!
I don't like a one-button help key...again, easy to accidently hit and start up an annoying help system. Better to use command-? or something.
Hmmm... but having a big button there is nice... back when I had to use a Sun5PC keyboard, before I got my Sun5Unix, I had a much larger area in which I could thwap my escape key. The way my keyboard layout is, I can hit escape, null, or help, and they all give me escape. I can sort of push my hand in that general area of the keyboard, and it will work... useful if I'm coming from an icky PC or Mac or something. Anyway, I miss the escape much less when it's above the tab then when it's an inch above the tilde. And I like having the backspace right next to the close bracket... pinky has to move less far.
Also, a few multiport adapter companies have converters for sun*PC* ... you could use that, but they're more expensive ($150+USD)
The nice thing about *nix is that you build the desktop system, you set up the user with a nice menu on the root window or on a button bar, and as long as that user doesn't know the root password, the user can't mess up the computer. That's why our PC group here has over 8x as many people as our unix group ... not becuase there are more PC's (the number is roughly the same, plus the servers are mostly UNIX), but because they are so much harder to maintain.
For the typical home user, or for small businiesses that just want a word processor or spreadsheet, Windows or MacOS is still probably their best choice -- the learning curve for Linux is probably not worth it (yet). But for a company that has to hire IT people and don't need windows-only programs, *nix can lower costs and problems a lot.
make sure you preserve permissions and things when using tar.
cp -av works nicely too (assuming gnu cp)
That's one of the things maintainers do: filter out all the crap code. That's why projects have maintainers. That's why we don't have a bazillion trojans in the code.
Now if only they give us some of the nice kernel multithreading that Irix has, and that 4dwm-zooming-filemanager...
for (i = 0; i < 100; i++)
int i,foo[100],baz[100]; /* Some stuff in here not dependant on */
/* Previous iterations */
/* We'll add foo and baz */
/* stuff in here */
for (i = 0; i
foo[i] += baz[i];
}
Now, that's silly. You don't want to do it for i=1, then for i=2, then for i=3, you just want to have it done 100 times, for values between 0 and 99. Adding two arrays, for instance, or zeroing them out. So, we need a construct such as
multido(i,0,99) {
/* that stuff in here */
foo[i] += baz[i];
}
Now, in most current and traditional program designs, this would still be non-trivial, as the compiler would have to start up two threads, let them process, wait for them to shut down, and so on, but in an architecture where overhead was less (pre-existing threads where you just pass code for them to run, maybe?) things like the above could be extremely beneficial.
Uh, didn't they say 6,000,000 shares? Doesn't that mean 60 million dollars worth? Or is the 6,000,000 just what they are releasing and doesn't accout for other investments?
The best part is you get those little tobasco sauce things.