Actually DOS Edit had this back when Apple was still pushing AppleWorks, which did not have this capability. Whether it came from one of the two companies is probably moot, as I imagine it came from a 3rd party editor first.
Actually Windowmaker is in the contributed section, as well as scores of other standard WMs that you don't know about (like CTWM). Enlightenment is somewhere in the publicly accessible area of afs also, but its not very freindly to the old versions of X that are run on andrew.
The best way is to look around and ask people (this goes for any school). I got my setup by first copying someone who I saw had a cooler setup than me, and asking them how I could get it. Later I started making it my own by configuring it how I wanted. Now, 3 other people run copies of my ctwm setup because they liked it better than what they had. If you look around you should see people with *Step; just ask them and they'll probably help you with it.
Disclaimer: I know ctwm is dead, but until any new WM is capable of what its been doing for 7 years, I'm not interested. (E looks like the best contender, but it still can't do some of the things ctwm can... go figure).
The reason Windows gets away with direct calls as a speed measure is that (1) Win has a broken differentiation between kernel and user space so switching among them is easier, and (2) they are on x86 so there isn't too much context in a contxt switch. What you get from pipes is free buffering so you don't have to do a context switch for every call (or two if you use safer user mode drivers). Believe it or not, windows NT was using a buffered design just like this before they decided to accept the instability of moving it into the kernel; and as it turns out the speed gains of the latter turned out to be minimal.
The main problems I see are people bitching incessantly about what they don't understand. I thought X was a dumb design until I hadd coded and worked on several different GUI systems. After all that most people realize (as did I) that X was a pretty wise tradeoff on all the design issues.
Apache seems to have a dumb design to me at this point; a lot of things don't seem to make sense; but that's probably due to my ignorance about its design issues.
It seems that evvery new security device at Airports is the answer to everything; metal detectors, X-ray scans of carryons, the bomb sniffers, etc.
What I'd like to know is how often these devices succesfully catch the criminals they are intended to detect. Of the past few hostage situations, did they pass through metal detectors with their guns? Are ceramic knives and guns being used to kill airline passangers all the time? My guess is *no*.
Of course the company that really doesn't have anything new to offer won't want to provide such numbers, would they? They just want the FAA to mandate its use, which should come in about 3 years...
Haven't done much in the way of network programming, have you? The interconnect is almost *always* slower than what can be done at the endpoints; hubble's downlink probably benefits a *lot* from doing compression and other processing onboard.
Besides, you can't control a sattelitte with ~2-5 second latency. Also, should the Hubble crash when their's an ion storm or other disturbasnce that breaks the communication with the base station? Think about it; something that expensive and that far away better have a high degree of autonomy...
If only speed matters, and not security, why on earth are they benchmarking SSL? Hello, anybody home?
Linux still has catching up to do to be on par with IIS in some types of static content; and we can again thank ZD for a meaningless benchmark that is no help in such development efforts.
Personally I can't wait to see what will happen with the 2.4 kernel with khttpd and phttpd on the horizon...
Remember when DOS/Windows had advantages over Linux? Yeah that was only at the *very* beginning, but that was the case for a short time during early development. Well here we are again.
Hurd has a higher degree of extensibility than traditional kernels. Since each individual feature is easier to add, eventually you may win out if your design is indeed better.
Supposing that people *only* work on improving things that already exist is short-sited. Everything you use today replaced something before it.
No, there *is* harm done, and there *is* a difference. If I ask for a service over the network, I'd expect that the server on the other end would be gathering statistics on that service. C'nom, who wouldn't? Same goes for stores that track customers through club cards.
But, if I'm not using a network service or buying from a store, I don't expect them to track me. If the store sends someone to follow me around after I leave their store, ICQ looks at files on my computer that I'm not using or sending, or RealCrapBox decides to send stuff over the network that I'm doing with information stored soley on my own machine -- I'm sorry, but these are unacceptable. Services should be monitored, but are not an invitation to monitor something beyond that service.
Hmm, hopefully it'll be good at drawing rather than just modifying existing pictures. IMHO Photoshop and Gimp are still behind Paint Shop and other programs for that sort of thing...
BTW, anyone know of any groups working on a sound editor?
I tried Dvorak, and found it easy to learn and much more comfotable when I was writing documents.
But... Using hard-coded or hard to change Alt-* and Ctl-* keys gets annoying. One example is Ctl-C, which is very useful to type while another hand is on a mouse, but impossible to do with the left hand without moving it to the right Ctl key. Not to mention that now various special keycodes can't be seen on the keyboard (the one case where almost everyone is hunk-and-peck).
My coding also slowed down due to the placement of keys for written text. I used to type with one hand on the editing keys (arrows, del, end, etc) and the other hand typing out keywords (Try it, it's pretty cool). The ping-pong back and forth typing with Dvorak made this nearly impossible to do easily with one hand.
So... Try it, esp. if you type words more than you type commands, and don't use inflexible programs that use a lot of mouse work. Then just pick what you like best...
By the way, let me thank the braindead Apple people for popularizing ZXCV for cut and paste, simply for their placement, making changing keymaps a royal pain in most OS's.
Take a classic board game and write your own computer version of it. Program "perfect" play for the computer player. Write a program to "solve" checkers through brute force. Write a fractal viewer with a cool zoom-in feature. Write a dense linear algebra package. Write a sparse linear algebra package. Get the edition of Numerical Recipes without the code and implement all the algorithms therein. Get the NR code and time test your implementations against theirs. Beat the times of the NR algorithms.
If you knew about game theory, NR, and linear algebra you must have had a pretty nice high school, and most likely it wasn't a standard public one. I was lucky to have an AP CS class at mine.
I think the point is that a lot of kids don't know what cool stuff is out there, like free development tools, Linux, and the like. That is changing, but much too slowly. We have to be a bit more active than osmisis...
"It is interesting to note that Dr. de Garis has made incredible progress by following a path the mainstream AI community has largely discounted -- that of modeling real neurons and real brain structures. I wonder what will come out of his next collaborative development at Starlab in Brussels? From his statements to me I would certainly hope he would find the living and working arrangements more congenial."
Umm, no. The reason people stopped trying this is that (1) we can't model everything about the neuron (2) what we did try didn't work (3) we don't know how real neurons learn.
This is probably a big backpropagation net on a chip, thus after 10,000 trials it will learn some stuff, while forgetting everything else that it learned before. If you ask connectionist people if the brain is a big set of backprop nets and nothing else, they will say "no" (notably among them would be McClelland).
The AIBO's sony sells adapt their behavior paramaters, but don't really learn. The modified AIBO's in Robocup had some learning. For example, the team from CMU (which I worked on) had a vision system that would learn in a limited way.
Machine learning right now depends mostly on the fact that problems are well broken up... Large scale, full "perception -> action" systems have so far been simplistic in what they learned, slow, or largely unsuccessful. I'll believe results, not speculations.
Hard, Sobering Facts: All they've evolved yet is some primitive motions in a simulator. Rodney Brooks & Co. did that on a real robot several years ago. It was by no means a trivial task.
It took Sony around 2 years to get a mobile quadruped working in the real world, after they already had a simulator for it in which it worked just fine.
Slashdot Mirror
Actually DOS Edit had this back when Apple was still pushing AppleWorks, which did not have this capability. Whether it came from one of the two companies is probably moot, as I imagine it came from a 3rd party editor first.
Actually Windowmaker is in the contributed section, as well as scores of other standard WMs that you don't know about (like CTWM). Enlightenment is somewhere in the publicly accessible area of afs also, but its not very freindly to the old versions of X that are run on andrew.
The best way is to look around and ask people (this goes for any school). I got my setup by first copying someone who I saw had a cooler setup than me, and asking them how I could get it. Later I started making it my own by configuring it how I wanted. Now, 3 other people run copies of my ctwm setup because they liked it better than what they had. If you look around you should see people with *Step; just ask them and they'll probably help you with it.
Disclaimer: I know ctwm is dead, but until any new WM is capable of what its been doing for 7 years, I'm not interested. (E looks like the best contender, but it still can't do some of the things ctwm can... go figure).
The reason Windows gets away with direct calls as a speed measure is that (1) Win has a broken differentiation between kernel and user space so switching among them is easier, and (2) they are on x86 so there isn't too much context in a contxt switch. What you get from pipes is free buffering so you don't have to do a context switch for every call (or two if you use safer user mode drivers). Believe it or not, windows NT was using a buffered design just like this before they decided to accept the instability of moving it into the kernel; and as it turns out the speed gains of the latter turned out to be minimal.
The main problems I see are people bitching incessantly about what they don't understand. I thought X was a dumb design until I hadd coded and worked on several different GUI systems. After all that most people realize (as did I) that X was a pretty wise tradeoff on all the design issues.
Apache seems to have a dumb design to me at this point; a lot of things don't seem to make sense; but that's probably due to my ignorance about its design issues.
It seems that evvery new security device at Airports is the answer to everything; metal detectors, X-ray scans of carryons, the bomb sniffers, etc.
What I'd like to know is how often these devices succesfully catch the criminals they are intended to detect. Of the past few hostage situations, did they pass through metal detectors with their guns? Are ceramic knives and guns being used to kill airline passangers all the time? My guess is *no*.
Of course the company that really doesn't have anything new to offer won't want to provide such numbers, would they? They just want the FAA to mandate its use, which should come in about 3 years...
Haven't done much in the way of network programming, have you? The interconnect is almost *always* slower than what can be done at the endpoints; hubble's downlink probably benefits a *lot* from doing compression and other processing onboard.
Besides, you can't control a sattelitte with ~2-5 second latency. Also, should the Hubble crash when their's an ion storm or other disturbasnce that breaks the communication with the base station? Think about it; something that expensive and that far away better have a high degree of autonomy...
If only speed matters, and not security, why on earth are they benchmarking SSL? Hello, anybody home?
Linux still has catching up to do to be on par with IIS in some types of static content; and we can again thank ZD for a meaningless benchmark that is no help in such development efforts.
Personally I can't wait to see what will happen with the 2.4 kernel with khttpd and phttpd on the horizon...
Remember when DOS/Windows had advantages over Linux? Yeah that was only at the *very* beginning, but that was the case for a short time during early development. Well here we are again.
Hurd has a higher degree of extensibility than traditional kernels. Since each individual feature is easier to add, eventually you may win out if your design is indeed better.
Supposing that people *only* work on improving things that already exist is short-sited. Everything you use today replaced something before it.
No, there *is* harm done, and there *is* a difference. If I ask for a service over the network, I'd expect that the server on the other end would be gathering statistics on that service. C'nom, who wouldn't? Same goes for stores that track customers through club cards.
But, if I'm not using a network service or buying from a store, I don't expect them to track me. If the store sends someone to follow me around after I leave their store, ICQ looks at files on my computer that I'm not using or sending, or RealCrapBox decides to send stuff over the network that I'm doing with information stored soley on my own machine -- I'm sorry, but these are unacceptable. Services should be monitored, but are not an invitation to monitor something beyond that service.
Make more sense now?
Hmm, hopefully it'll be good at drawing rather than just modifying existing pictures. IMHO Photoshop and Gimp are still behind Paint Shop and other programs for that sort of thing...
BTW, anyone know of any groups working on a sound editor?
I tried Dvorak, and found it easy to learn and much more comfotable when I was writing documents.
But...
Using hard-coded or hard to change Alt-* and Ctl-* keys gets annoying. One example is Ctl-C, which is very useful to type while another hand is on a mouse, but impossible to do with the left hand without moving it to the right Ctl key. Not to mention that now various special keycodes can't be seen on the keyboard (the one case where almost everyone is hunk-and-peck).
My coding also slowed down due to the placement of keys for written text. I used to type with one hand on the editing keys (arrows, del, end, etc) and the other hand typing out keywords (Try it, it's pretty cool). The ping-pong back and forth typing with Dvorak made this nearly impossible to do easily with one hand.
So...
Try it, esp. if you type words more than you type commands, and don't use inflexible programs that use a lot of mouse work. Then just pick what you like best...
By the way, let me thank the braindead Apple people for popularizing ZXCV for cut and paste, simply for their placement, making changing keymaps a royal pain in most OS's.
Take a classic board game and write your own computer version of it. Program "perfect" play for the computer player. Write a program to "solve" checkers through brute force. Write a fractal viewer with a cool zoom-in feature. Write a dense linear algebra package. Write a sparse linear algebra package. Get the edition of Numerical Recipes without the code and implement all the algorithms therein. Get the NR code and time test your implementations against theirs. Beat the times of the NR algorithms.
If you knew about game theory, NR, and linear algebra you must have had a pretty nice high school, and most likely it wasn't a standard public one. I was lucky to have an AP CS class at mine.
I think the point is that a lot of kids don't know what cool stuff is out there, like free development tools, Linux, and the like. That is changing, but much too slowly. We have to be a bit more active than osmisis...
An ex-bored teenager...
"It is interesting to note that Dr. de Garis has made incredible progress by following a path the mainstream AI community has largely discounted -- that of modeling real neurons and real brain structures. I wonder what will come out of his next collaborative development at Starlab in Brussels? From his statements to me I would certainly hope he would find the living and working arrangements more congenial."
Umm, no. The reason people stopped trying this is that (1) we can't model everything about the neuron (2) what we did try didn't work (3) we don't know how real neurons learn.
This is probably a big backpropagation net on a chip, thus after 10,000 trials it will learn some stuff, while forgetting everything else that it learned before. If you ask connectionist people if the brain is a big set of backprop nets and nothing else, they will say "no" (notably among them would be McClelland).
The AIBO's sony sells adapt their behavior paramaters, but don't really learn. The modified AIBO's in Robocup had some learning. For example, the team from CMU (which I worked on) had a vision system that would learn in a limited way.
Machine learning right now depends mostly on the fact that problems are well broken up... Large scale, full "perception -> action" systems have so far been simplistic in what they learned, slow, or largely unsuccessful. I'll believe results, not speculations.
Hard, Sobering Facts:
All they've evolved yet is some primitive motions in a simulator. Rodney Brooks & Co. did that on a real robot several years ago. It was by no means a trivial task.
It took Sony around 2 years to get a mobile quadruped working in the real world, after they already had a simulator for it in which it worked just fine.