While I agree with most of what you said, one problem with the car and computer analogy is that people who own modern cars, as refined a technology as they are, typically bring them in for service at least four times a year if they aren't able to do their own oil changes and other routine maintainance. I don't think most home users use any kind of professional servicing on their computers. Imagine how different the experience would be for most home users if they brought in their computer every two months to have their computers cleaned up! (E.g., fix configuration problems, remove spyware, apply patches, defrag disk, etc)
For a closer analogy to the typical new car owner, imagine if users didn't know their own administrative password (or at least it was buried in some manual that comes with the computer) and had to bring in their machine to make any configuration changes? (Okay, this method would be much easier on OS X than Windows...) Right now most home computer users are like the guy who tries to work on his car but just screws things up.
Having to spend four hours on hold to get tech support and then go through a bunch of steps that probably won't work over the phone discourages people from even calling help phone lines.
I think one key difference is that people need their cars to be running well for both personally safety and to be able to get to work, so if their car is belching smoke or the steering wheel is wobbling around they (usually) will take it in to a mechanic, whereas a person who's home computer is full of spyware, popups, error messages, and random crashes will probably just limp along for months or try to get some friend to fix it. Or not. For all this bitching about techies having to do free tech support, haven't you ever tried to use a friend or relative's computer and it was so messed up that you felt compelled to try and fix it without being asked?
Just as another comparison to cars, how much time do you spend crapping around with your car versus hours actually driving it? I use my work computer probably 8 hours a day, but my car less than two. For the car (non power-user), one has to include pumping gas, washing the car, filling tires and washer fluid, changing bulbs, dealing with paperwork (insurance and registration), scheduling service appointments, driving to and from service center, hanging around in the service center, etc, etc. If you own a used car and/or do some of your own work, than you have to add many more hours of crapping around with all kinds of stuff.
I think time spent messing around with computers is often more annoying because most of the "car servicing" time is fairly periodic and expected, whereas usually computers fail right when you're trying to do something important. It's rare that a car will fail catastrophically; usually most problems can be deferred for a few days (except for a flat tire, but changing a tire takes less time than reinstalling windows!) Furthermore, if your car does die, you can rent a car and start using it immediately, whereas getting a new computer to the point where you can get your work done is usually more involved.
I'd like to hear if anyone else has had changes in migraine patterns after quitting coffee. It would be really hard to give up "the bean", but not having migraines would be a good motivating factor...
A funny book for anyone who thinks about coffee is Memoir from Antproof Case by Mark Helprin--the protagonist has a lifelong semi-irrational hatred of coffee and "coffee apologists".
Now, if you said you built a laser Microphone with 90 dB directional gain and pointed it at the guys window so you could listen to the music he plays, you'd be ok.
Well, that's not a very good analogy either. By tapping into this business' wireless network, he is using up bandwidth that the business is paying for. Imagine the students are continuously downloading pr0n while the business is trying to download some mission-critical data. Time is money.
Pointing a microphone at his window and listening in on whatever music the guy happens to be playing is passive sensing, perhaps an invasion of privacy but doesn't affect anything in the house. Using a wifi is a two-way transaction--he is sending packets to the business' WAP and forcing it to steal bandwidth from the business. I don't know if the legal system sees that the same as physically entering a building, but I would guess if you robbed a store remotely using magnets the law would find something...
Perhaps a better analogy would be constructing a really powerful remote control that can operate this guy's stereo system and using that in addition to the laser microphone. If they only used it while the guy wasn't in his room, they would be just stealing electricity and adding wear and tear on his system, but if they used it while he was trying to listen to something else, that would be more of an inconvenience. While it's quite possible that Herr Professor is only stealing a small enough chunk of bandwidth that the business will never notice, it would be easy even accidentally to use up a bunch of bandwidth. Someone could probably sneak into my house and steal a few raisins every day out of my pantry and I wouldn't notice either.
I recently found an open second-story window (ie., no working lock) near the campus where I teach music theory.
In order to allow my students to listen to music, and because our equipment budget is small and hence we don't have stereo equipment in classrooms, I was forced to, along with several students, carry out an activity in the middle of the pitch black night in the name of education.
That's right -- myself and several students built a ladder and climbed into the window (in a studio nearby) to steal some Tannoy reference speakers and installed them in my third floor office/teaching area.
The students can now freely enjoy high quality sound reproduction, and I for one won't budge. This is no theft, it's all done in the name of education.
I dare somebody to sue me. Oh, and by the way... we use tube amplifiers.
The problem I have noticed is that it often takes more mental energy to set up a macro or whatever to do something rather than just blindly do something by hand. I'll often find my coworkers who use CAD programs doing these incredibly repetitive tasks (probably causing repetitive stress injury) that could be done with macros, and while sometimes it's because that particular task is hard to do with a macro, often it's just because it's easier just to start clicking than to have to plan out a macro, even if it will save time. Often if your work involves a lot of problem solving and making decisions, having a small burst of repetive/mindless "busywork" can actually be sort of a break. Obviously after a certain point it becomes really annoying if there's no good way to automate that process. I think different people get annoyed at different points--a lot of programmers will write a macro to do something two times!
I've caught myself doing the same thing in the code editor--I've tried to make myself always quickly define a temporary macro to do those things just to save wear and tear on my tendons.
The problem with repetitive tasks is that you're typing at maximum speed for an extended period of time, while in more typical coding there's usually a few pauses for thinking or compilation.
To get vaguely back on topic, one case where I find a CLI infinitely superior to GUI is operations involving displaying, moving, deleting, or copying a lot of files but only certain kinds. E.g. in Windows,
del source\*.cpp
from the CLI is so much faster than opening the folder window, selecting "Detail" view, clicking on the "File type" sorting tab, scrolling down and finding the _first_.cpp file (easy to be off by one), clicking on it, scrolling down to the last one, and then shift-clicking on the last one (and hoping you didn't accidentally click on a file while scrolling down), right-click and choose "Delete".
This is a case that probably happens more to coders than other people, but I'm very curious to see what an OS that treated "folder windows" as more like database query results would do for those types of operations, especially if it was very easy to quickly create a new custom view, use it, and discard it (kind of like smart playlists in iTunes). Supposedly longhorn will attempt some of this stuff, but presenting a good interface to this sort of power (that non-coders can use) is not trivial. When dealing with files, I often switch back and forth between doing operations in the folder windows and doing them from the command line depending on the specific task, but it would be nice if both methods could employ more of the power of the other (e.g., being able to easily see "property" and "summary" information from the CLI.)
OSX has a Darwin/MACH kernel with a NeXTSTEP-derived application layer; the BSD bits are mostly just optional command-line utilities. Also, I believe that Win2K is internally NT 5.0, and XP is NT 5.1 (I think an actual NT 5.0 was planned, but it was so late it morphed into Win2K) </pedantic>
I think one could argue that OS9 and OSX are so different it would be like Microsoft taking FreeBSD or Linux, heavily modifying Gnome, writting a very comprehensive/official version of Wine for backwards compatibility, and releasing the whole package as the next version of Windows.
A human brain has roughly 10 billion neurons, with complex interconnections, IIRC. While it takes a long time for a neuron to react compared to a modern transistor, brains process many things at the same time, and without the computer's clear-cut distinction between memory and processor. A PC might have 10 billion transistors in its RAM, but they mostly just sit there until the CPU has a cache miss, which many of our neurons can be "working" at once. For a neural net computer, it might be possible to come up with some sort of metric to compare neurons to virtual "neurodes", but a method to compare Pentium clock cycles to neuron firing times would be pretty subjective.
In fact, it is this fundamental difference in how humans and (conventional) computers work that makes the man-vs-machine issue interesting. (E.g., a slow, massively parallel network built for collecting grubs and and having sex, against a fast, serial turing machine built for number crunching)
If you're displaying a 24 fps movie on a CRT running at 72 Hz, there are three refreshes per frame. Each refresh on a CRT is a transition between a dark blank and a bright image, which is a noticable high contrast change, whereas on an electronic ink method it could just be a transition from one bright frame to another, with a high degree of coherency between the two images. (This is assuming that the intermediate state of an ink pixel while changing colors lies somwhere between the old color and new color, which seems likely.)
For CRT refresh rates, it seems like 60 Hz is about the minimum for a display in the center of your vision (at least to my NTSC-tuned eyes, PAL looks noticably flickery), and for peripheral vision, it seems like about 75 Hz is required. However, for *frame* rates, it seems like 24 fps is acceptable to most people (e.g., movies), although I think that while 30 fps for CG looks "smooth", 60 fps has this additional "glasslike" quality. The worse cases are rapid pans or trains moving past the screen horizontally. There may be an argument that temporal antialiasing could make a 30 fps rendering look almost as good as 60 fps.
I also wonder whether a true 24 Hz device would display a 24 fps movie better than an interlaced 60 Hz device due to the annoying 3:2 "beat" pattern you can see on NTSC conversions of movies.
Re:Looking forward... mostly
on
Quicksilver
·
· Score: 1
The complete misunderstanding of Chomsky in Snow Crash was also pretty grating--it read like intro linguistics as misunderstood by an air-headed freshman psych major. Cryptonomicon was also full of historical innaccuracies and misrepresentations. I've noticed tons of false statements in areas that I happen to know a little about, so I have to extrapolate that there is a lot of goofs in areas I know nothing about. While I enjoyed reading it, I hate the thought that I've unwittingly encorporated a bunch of misinformation into my head.
Even though it is fiction, there is sort of an unspoken agreement that statements that are presented as factual in fiction (especially fiction that has a historical or scientific theme) should really be true. I.e., it's okay to make your characters do anything, or make up stuff that historical characters might have done, but don't distort the world when speaking as omniscient narrator, especially if it is due to just not caring about the facts instead of making a concession to the plot of the story.
I think Stephenson is just sort of sloppy and lazy with his research; from reading over "the Codebreakers", it looks like he may have skimmed through a few chapters and then wrote a zillion words. He just doesn't seem to feel like doing the homework is worthwhile, and his politics seem a bit naiive. He also could use to trim most of his stuff by 50% to get a better signal-to-noise ratio.
It's a pity, though, because he is very inventive in many ways, and some of his portrayals of how geeks view the world in Cryptonomicon are priceless. He also has a great sense of delving for interesting topics, times, and places; maybe he just needs to hire a few fact-checkers to do the legwork... or at least maybe some kind souls could start an official Stephenson debunking/debriefing website where we could go to purify ourselves after reading...
I guess at some point, who really cares if the code to swap ctrl and caps is a kludge or not as long as it works? For the record, I'm writing this from an older TiBook with uControl, but the caps-lock light still toggles on and off even though it comes through as ctrl, which is lame, but definitely quite usable for emacs. (My guess is that uControl must be getting "cooked" characters and just swapping the case if the capslock key is down and then synthesizing the control key? If so, that's really nasty, but it's the kind of thing that you can just forget about.) If nobody has written a keyswapper for *BSD market, that's a good clue that the number of users is too small for Apple to care about.
No matter how screwed up the hardware is, if it can be remapped in some OS's, it has to be possible to remap it in OpenBSD somehow, and it should also be possible to hack in something to get some function keys to show up as alternate mouse buttons.
Since OS-X already supports X, BSD utilities, and POSIX, it's pretty clear that Apple must view people running other unices on their hardware as a teeny minority. They are doing a good job of wooing linux users over to OS-X by writing X-windows implementations and so forth, so why would they want to encourage people to run OpenBSD on their laptops, which can't run iTunes et al. Plus, when *nix people have to run OS-X on their Mac laptops, this will make it more likely that they'll just port their stuff to work with OS-X, which will than benefit "regular" OS-X users, and thus help out Apple by providing more software that can run on their native OS. Someone who buys a Powerbook and installs *nix on it is someone who could easily switch to a Thinkpad next year; someone who buys a Powerbook and runs OS-X will likely get "hooked" and want to buy another Powerbook.
Marketing schemes aside, it's even more likely that the non-MacOS users of Mac hardware are such a small minority that Apple doesn't even spend much time thinking about them. As a software developer, I've often had users with special needs ask for all kinds of features, not realizing that they are.003% of the market or whatever, and supporting their esoteric requests is more goodwill than actual good business practice. Goodwill is important for sure, but at the end of the day there's only so much you can work on and you have to use some sort of triage.
Here's a strange idea--an overachieving RSV could reach a point that could intersect a low earth orbit. Theoretically, an RSV could spit out a payload that would lasso an orbiting payload of the same mass as it passed by, and slingshot the new payload into orbit and bring the old payload down to suborbital velocities, where it could either be stored in the RSV or just allowed to reenter the atmosphere on its own.
The forces on the "lasso" would be extremely high, quite possibly beyond current (or even any) material, although making the lasso quite long would spread out the kinetic energy exchange over a longer period of time, reducing the instantaneous accelerations on the payloads.
One possible application would be sending up payloads consisting of a booster rocket plus equipment; the payload would travel to say a lunar mining station, and the returning payloads would be mineral resources. If a similar system were used on the lunar end, in theory the only energy expended would be to compensate for frictional losses (plus the energy to "prime the pump" with the first payload); getting out of the gravity well would be "free". (Much like a hybrid Civic with regenerative brakes stopping and starting up again)
Of course, IANARS, but I would love to hear why this wouldn't work...
While debates about which local dialect is the "correct" or "true" dialect are no more scientific than phrenology, (as in the old joke "a language is a dialect with an army and a navy") another amusing American/British difference is the word aluminum (aluminium). The full story is below.
Am I the only one who felt the cold hand of mortality when I saw that reference to an "ancient Motorola 68k"? I remember when the 68030 was one mean machine... although I admit we flipped switches attached to a 6502, which isn't vastly older...
(the truly "ancient" machine at my school was the old PDP8 that we used in one class by entering hand-assembled programs in with flip switches)
I'm not "using" my copies of 10.0 and 10.1!
on
Jaguar is Over
·
· Score: 1
Well, I'm hardly "using" the OS security updates on my "copies" of 10.0 and 10.1, since I'm running 10.2 now. By linking the more quickly changing media technologies to a paid OS point release, plus releasing 1st party apps that only run on the latest version, one feels much more "forced" to always have the latest Mac OS, whereas in Windows you have the independent DirectX and IE releases that let you have a 3-4 year old OS and not feel penalized, even for running games and other "bleeding edge" technology.
Granted, Apple must amortize their OS development costs over a much smaller user base, and so one expects to pay somewhat more for Mac OS than Windows (although they do have the closed hardware platform as another revenue stream, as well as a vastly reduced set of hardware to test against). However, it would be still nice to get even a nominal discount from Apple for upgrading one's OS annually. Given that most rebate programs tend to get less than 10% participation (shh!!), a $15 rebate would make people "feel good" and probably not cost Apple much at all...
Economically, probably those of us who upgrade annually are likely to do it anyways, so Apple is probably making the right choice financially to soak us, but they could at least whitewash it a little more so we can continue to lie under the pleasant thrum of the Steve Jobs reality distortion field while we eat our food pellets and the AIBO mows the chem-lawn outside...
As if the world needs yet another opinion on this, but feel like this might be useful to techy wintel users who are thinking of switching...
I've used wintel since DOS days, and use Win2K at work, but my current home computer is a PowerBook G4 (the hardware looked cool, plus I was getting irritated at MS). While it took a while to get fully comfortable with OS-X, I have the feeling that people with Windows and *NIX experience may actually like OS-X better than do long-time Mac users...
Anyways, I like OSX in many ways, but like all OS's it has strengths and weaknesses. On the plus side, the whole package (hardware and software) is very visually pleasing, and it's clear they take interface design seriously. Basically, it feels like it was designed by designers instead of engineers, with both the advantages and disadvantages you might expect of that. Most of their apps sacrifice a large feature set for a tight and clean app that does everything you want 99% of the time. The NeXT-derived method of packaging applications into a single icon that you can drag anywhere, use, and then drag off with no installation is really the right way of doing things. The iApps are really nice. Being able to bring up a unix shell window with gcc and emacs and all that is very nice, as well.
On the minus side, the machine is not a speed demon no matter what the apologists say. While sure, not many apps are optimized for altivec, but then again how many apps are optimized for MMX and SSE(n)? For most of my modest home needs, however, it has more than enough power. However, the OS doesn't feel as "real time" as it should--often clicking on something will bring up the "spinning wheel" cursor for several seconds or occasionally much more before a menu pops up. Probably adding more RAM (I have 256 currently) would address this, but it feels like something in their windowing system is blocking that shouldn't be. Also some lack of polish such as the cursor sometimes showing the wrong thing (e.g., I-beam when over a button), Finder windows occasionally defaulting to messed-up column widths, etc, that have been there through several updates is a little sketchy given that they control the entire hardware platform... another issue is that a lot of open-source stuff seems to be hard to get to build under OSX/Darwin, although this will probably improve over time. You can definitely feel that they have a lot fewer programmers than MS, but this seems to cut both ways.
I sort of resented having to pay to upgrade to 10.2 to get decent audio support, and there are still apps that haven't been ported to OS-X natively (and running classic mode feels much less seamless than running DOS and Win16 apps did under Win95) Microsoft at least made their fast-changing multimedia layer (including DirectX) seperate and freely upgradable independent of the OS, so you could run a lot of new stuff on '95 for quite a few years--with Mac it seems like you have to commit to upgrading every year or so to be able to run most programs. Also, the nice built-in apps make (brainwash?) you want everything else on your system to work/look the same (I didn't have the same problem with Office for Windows for some strange reason...;)
I don't feel like I'm getting the absolute most "bang for my buck", but I've gotten more and more sick of fscking around with malfunctioning computers, especially in precious free time, that I'm willing to pay a premium for a pretty and reliable system. However, I still like my Win2K box at work just fine as well. I think switching between OS's every so often is a good idea for anybody just to keep from getting to attached to one paradigm...
Why not just get an optical PC-CARD or USB adapter and go digital out straight into your receiver? Those are pretty cheap, and since it's digital there theoretically shouldn't be any loss. Since you're thinking of springing for an HDTV plasma display, I'm assuming your receiver can do Dolby Digital or DTS 5.1 already and should have decent D-A converter on it.
Re:Do they still make "demos"?
on
The Cg Tutorial
·
· Score: 2, Interesting
Modern PC graphics hardware gotten to the point that there is much less emphasis on clever hackery, and more on just generating good source art. In the old demos, the general-purpose CPUs with simple bit-mapped video cards weren't powerful enough to do a lot of 3D effects, but clever programming could 'fake it out' in various ways, usually by taking advantage of some special case (such as the "2.5D" games like Wolfenstein3D). Nowadays, we have specialized 3D hardware pipelines that solve the general/correct case for most standard graphics, and it becomes more a matter of stringing together various primitive operations to get an asthetically-pleasing effect, so it's more "artistic" and less "clever" programming.
I think a console like the PS2 is a much better place for exhibiting the old-fashioned "clever hack", because its weird and exposed architecture allows for many, many more ways to do things, some of which can be very "crazy". Plus hand-coding assembly to fit in 4K or 16K for the vector units seems very similar to the old demo contests... is there a demo scene for the PS2 linux kit? There should be if there isn't...
It seems like either way you have to "know about" every leap second--if you use TAI then you can subtract times and get correct intervals, but to display a time you have to do a conversion. If you use UTC, then you have to deal with the fact that time doesn't advance uniformly.
I like the fact that NT's 64-bit time counts something like 100s of microseconds since 1600 AD instead of the old 32-bit DOS-style "seconds since 1970". This has the nice property that it's already larger than 4 billion, so it makes it harder to screw up an implementation without seeing an error immediately. Perhaps if they made TAI not even close to UTC then it would be easier to spot bugs. Perhaps you could even just store both values in timestamps, and use TAI for computation/timestamping and UTC for display... that way you just depend on the OS to keep both values up to date, and less room for application bugs...
...requires coding explicitly for that platform. That's one "secret" of why consoles seem to have better graphics performance than a PC with similar specs. While someone could probably write a good OpenGL driver for the Xbox (due to its similarities to a GF3), you would need inside information or have to do a lot of reverse engineering. The GameCube would probably be a similar case--its GPU would probably be a pretty good match with OpenGL.
However, I don't think it would be possible to get a generic OpenGL implementation to run well on a PS2, even with inside information--with its 300 MHz CPU and no L2 cache, single-texturing, 4MB VRAM, no hardware clipping, etc, it would perform very poorly on platform-agnostic code and data. It would also be very slow to convert 32-bit textures into palettized on the fly unless you did a really crappy job of it. The PS2 will perform poorly unless you carefully structure your game around it's unique needs, and memory will be very scarce if your driver has to copy and translate your geometry and textures into a hardware-friendly format. PC-based graphics cards are designed around running DirectX and OpenGL efficiently--the PS2 has no such requirement. You might have a chance for some performance at the expense of memory if the game drew huge portions of the world into a display list, allowing the driver to convert this into a big DMA chain.
You've all seen how the first PS2 games often looked worse than Dreamcast games, and these were from professional developers who were coding for the PS2!
I suppose if you just wanted to run Minesweeper or something you'd be okay...
The "wheels shoved out to the sides" cars I've seen were most definitely cars that had never seen a race track, more of a style thing--it's not the usual "rice" thing, because they don't seem to have extra spoilers and "Type R" stickers.
Anyways, I myself went for camber plates instead of bending my struts, although the guy I know who went the bending route has had tons of track time on that car with no problems. Since at the present time I just do open lapping sessions with no racing, I'm much more interested in mods that make the handling more neutral, keep the brakes from fading, or make the car more reliable under track conditions, rather than eaking out a little more horsepower or shaving off a few pounds.
It's interesting that the performance-oriented wing of the Japanese import crowd seems much more focused on drag racing, while their counterparts in the Bimmer/Porsche groups go for road racing as the yardstick by which to judge a sports car. Different strokes for different folks, I guess.
(or worse, suspensions lowered simply by cutting the springs already there, rather than replacing with proper springs)
Actually, slightly lowering your car by cutting off a half-coil is recommended by some respected motorsports books because it not only reduces the center of gravity but slightly increases the effective spring-rate as well. I also know people who've bent their front struts to gain some negative camber, but that seems a little sketchy to me...
I agree, though, that most of these modded cars are slower than stock, although if I see one with big floppy drag tires on the front wheels I have to allow for the possibility of it being fast. My other favorite is the cars whose main modification seems to be sticking the wheels out an extra six inches or so--the thought of all that shearing force on the bearings makes me cringe...
Slashdot Mirror
For a closer analogy to the typical new car owner, imagine if users didn't know their own administrative password (or at least it was buried in some manual that comes with the computer) and had to bring in their machine to make any configuration changes? (Okay, this method would be much easier on OS X than Windows...) Right now most home computer users are like the guy who tries to work on his car but just screws things up.
Having to spend four hours on hold to get tech support and then go through a bunch of steps that probably won't work over the phone discourages people from even calling help phone lines.
I think one key difference is that people need their cars to be running well for both personally safety and to be able to get to work, so if their car is belching smoke or the steering wheel is wobbling around they (usually) will take it in to a mechanic, whereas a person who's home computer is full of spyware, popups, error messages, and random crashes will probably just limp along for months or try to get some friend to fix it. Or not. For all this bitching about techies having to do free tech support, haven't you ever tried to use a friend or relative's computer and it was so messed up that you felt compelled to try and fix it without being asked?
Just as another comparison to cars, how much time do you spend crapping around with your car versus hours actually driving it? I use my work computer probably 8 hours a day, but my car less than two. For the car (non power-user), one has to include pumping gas, washing the car, filling tires and washer fluid, changing bulbs, dealing with paperwork (insurance and registration), scheduling service appointments, driving to and from service center, hanging around in the service center, etc, etc. If you own a used car and/or do some of your own work, than you have to add many more hours of crapping around with all kinds of stuff.
I think time spent messing around with computers is often more annoying because most of the "car servicing" time is fairly periodic and expected, whereas usually computers fail right when you're trying to do something important. It's rare that a car will fail catastrophically; usually most problems can be deferred for a few days (except for a flat tire, but changing a tire takes less time than reinstalling windows!) Furthermore, if your car does die, you can rent a car and start using it immediately, whereas getting a new computer to the point where you can get your work done is usually more involved.
A funny book for anyone who thinks about coffee is Memoir from Antproof Case by Mark Helprin--the protagonist has a lifelong semi-irrational hatred of coffee and "coffee apologists".
2. ???
3. Profit!
Pointing a microphone at his window and listening in on whatever music the guy happens to be playing is passive sensing, perhaps an invasion of privacy but doesn't affect anything in the house. Using a wifi is a two-way transaction--he is sending packets to the business' WAP and forcing it to steal bandwidth from the business. I don't know if the legal system sees that the same as physically entering a building, but I would guess if you robbed a store remotely using magnets the law would find something...
Perhaps a better analogy would be constructing a really powerful remote control that can operate this guy's stereo system and using that in addition to the laser microphone. If they only used it while the guy wasn't in his room, they would be just stealing electricity and adding wear and tear on his system, but if they used it while he was trying to listen to something else, that would be more of an inconvenience. While it's quite possible that Herr Professor is only stealing a small enough chunk of bandwidth that the business will never notice, it would be easy even accidentally to use up a bunch of bandwidth. Someone could probably sneak into my house and steal a few raisins every day out of my pantry and I wouldn't notice either.
In order to allow my students to listen to music, and because our equipment budget is small and hence we don't have stereo equipment in classrooms, I was forced to, along with several students, carry out an activity in the middle of the pitch black night in the name of education.
That's right -- myself and several students built a ladder and climbed into the window (in a studio nearby) to steal some Tannoy reference speakers and installed them in my third floor office/teaching area.
The students can now freely enjoy high quality sound reproduction, and I for one won't budge. This is no theft, it's all done in the name of education.
I dare somebody to sue me. Oh, and by the way... we use tube amplifiers.
(Sorry, I just couldn't stop myself!)
I've caught myself doing the same thing in the code editor--I've tried to make myself always quickly define a temporary macro to do those things just to save wear and tear on my tendons. The problem with repetitive tasks is that you're typing at maximum speed for an extended period of time, while in more typical coding there's usually a few pauses for thinking or compilation.
To get vaguely back on topic, one case where I find a CLI infinitely superior to GUI is operations involving displaying, moving, deleting, or copying a lot of files but only certain kinds. E.g. in Windows,
from the CLI is so much faster than opening the folder window, selecting "Detail" view, clicking on the "File type" sorting tab, scrolling down and finding the _first_This is a case that probably happens more to coders than other people, but I'm very curious to see what an OS that treated "folder windows" as more like database query results would do for those types of operations, especially if it was very easy to quickly create a new custom view, use it, and discard it (kind of like smart playlists in iTunes). Supposedly longhorn will attempt some of this stuff, but presenting a good interface to this sort of power (that non-coders can use) is not trivial. When dealing with files, I often switch back and forth between doing operations in the folder windows and doing them from the command line depending on the specific task, but it would be nice if both methods could employ more of the power of the other (e.g., being able to easily see "property" and "summary" information from the CLI.)
OSX has a Darwin/MACH kernel with a NeXTSTEP-derived application layer; the BSD bits are mostly just optional command-line utilities.
Also, I believe that Win2K is internally NT 5.0, and XP is NT 5.1 (I think an actual NT 5.0 was planned, but it was so late it morphed into Win2K)
</pedantic>
I think one could argue that OS9 and OSX are so different it would be like Microsoft taking FreeBSD or Linux, heavily modifying Gnome, writting a very comprehensive/official version of Wine for backwards compatibility, and releasing the whole package as the next version of Windows.
In fact, it is this fundamental difference in how humans and (conventional) computers work that makes the man-vs-machine issue interesting. (E.g., a slow, massively parallel network built for collecting grubs and and having sex, against a fast, serial turing machine built for number crunching)
It stands for Storage Area Network. Here's a link courtesy of our overlords at Google.
For CRT refresh rates, it seems like 60 Hz is about the minimum for a display in the center of your vision (at least to my NTSC-tuned eyes, PAL looks noticably flickery), and for peripheral vision, it seems like about 75 Hz is required. However, for *frame* rates, it seems like 24 fps is acceptable to most people (e.g., movies), although I think that while 30 fps for CG looks "smooth", 60 fps has this additional "glasslike" quality. The worse cases are rapid pans or trains moving past the screen horizontally. There may be an argument that temporal antialiasing could make a 30 fps rendering look almost as good as 60 fps.
I also wonder whether a true 24 Hz device would display a 24 fps movie better than an interlaced 60 Hz device due to the annoying 3:2 "beat" pattern you can see on NTSC conversions of movies.
Even though it is fiction, there is sort of an unspoken agreement that statements that are presented as factual in fiction (especially fiction that has a historical or scientific theme) should really be true. I.e., it's okay to make your characters do anything, or make up stuff that historical characters might have done, but don't distort the world when speaking as omniscient narrator, especially if it is due to just not caring about the facts instead of making a concession to the plot of the story.
I think Stephenson is just sort of sloppy and lazy with his research; from reading over "the Codebreakers", it looks like he may have skimmed through a few chapters and then wrote a zillion words. He just doesn't seem to feel like doing the homework is worthwhile, and his politics seem a bit naiive. He also could use to trim most of his stuff by 50% to get a better signal-to-noise ratio.
It's a pity, though, because he is very inventive in many ways, and some of his portrayals of how geeks view the world in Cryptonomicon are priceless. He also has a great sense of delving for interesting topics, times, and places; maybe he just needs to hire a few fact-checkers to do the legwork... or at least maybe some kind souls could start an official Stephenson debunking/debriefing website where we could go to purify ourselves after reading...
No matter how screwed up the hardware is, if it can be remapped in some OS's, it has to be possible to remap it in OpenBSD somehow, and it should also be possible to hack in something to get some function keys to show up as alternate mouse buttons.
Since OS-X already supports X, BSD utilities, and POSIX, it's pretty clear that Apple must view people running other unices on their hardware as a teeny minority. They are doing a good job of wooing linux users over to OS-X by writing X-windows implementations and so forth, so why would they want to encourage people to run OpenBSD on their laptops, which can't run iTunes et al. Plus, when *nix people have to run OS-X on their Mac laptops, this will make it more likely that they'll just port their stuff to work with OS-X, which will than benefit "regular" OS-X users, and thus help out Apple by providing more software that can run on their native OS. Someone who buys a Powerbook and installs *nix on it is someone who could easily switch to a Thinkpad next year; someone who buys a Powerbook and runs OS-X will likely get "hooked" and want to buy another Powerbook.
Marketing schemes aside, it's even more likely that the non-MacOS users of Mac hardware are such a small minority that Apple doesn't even spend much time thinking about them. As a software developer, I've often had users with special needs ask for all kinds of features, not realizing that they are .003% of the market or whatever, and supporting their esoteric requests is more goodwill than actual good business practice. Goodwill is important for sure, but at the end of the day there's only so much you can work on and you have to use some sort of triage.
The forces on the "lasso" would be extremely high, quite possibly beyond current (or even any) material, although making the lasso quite long would spread out the kinetic energy exchange over a longer period of time, reducing the instantaneous accelerations on the payloads.
One possible application would be sending up payloads consisting of a booster rocket plus equipment; the payload would travel to say a lunar mining station, and the returning payloads would be mineral resources. If a similar system were used on the lunar end, in theory the only energy expended would be to compensate for frictional losses (plus the energy to "prime the pump" with the first payload); getting out of the gravity well would be "free". (Much like a hybrid Civic with regenerative brakes stopping and starting up again)
Of course, IANARS, but I would love to hear why this wouldn't work...
Although to retain that critical "slashdot bloc", Dean should probably also make some sort of public apology, perhaps via another mass e-mailing.
Element Al
(the truly "ancient" machine at my school was the old PDP8 that we used in one class by entering hand-assembled programs in with flip switches)
Granted, Apple must amortize their OS development costs over a much smaller user base, and so one expects to pay somewhat more for Mac OS than Windows (although they do have the closed hardware platform as another revenue stream, as well as a vastly reduced set of hardware to test against). However, it would be still nice to get even a nominal discount from Apple for upgrading one's OS annually. Given that most rebate programs tend to get less than 10% participation (shh!!), a $15 rebate would make people "feel good" and probably not cost Apple much at all...
Economically, probably those of us who upgrade annually are likely to do it anyways, so Apple is probably making the right choice financially to soak us, but they could at least whitewash it a little more so we can continue to lie under the pleasant thrum of the Steve Jobs reality distortion field while we eat our food pellets and the AIBO mows the chem-lawn outside...
2000 * 5280 / 150 =
70,400 volunteers
I've used wintel since DOS days, and use Win2K at work, but my current home computer is a PowerBook G4 (the hardware looked cool, plus I was getting irritated at MS). While it took a while to get fully comfortable with OS-X, I have the feeling that people with Windows and *NIX experience may actually like OS-X better than do long-time Mac users...
Anyways, I like OSX in many ways, but like all OS's it has strengths and weaknesses. On the plus side, the whole package (hardware and software) is very visually pleasing, and it's clear they take interface design seriously. Basically, it feels like it was designed by designers instead of engineers, with both the advantages and disadvantages you might expect of that. Most of their apps sacrifice a large feature set for a tight and clean app that does everything you want 99% of the time. The NeXT-derived method of packaging applications into a single icon that you can drag anywhere, use, and then drag off with no installation is really the right way of doing things. The iApps are really nice. Being able to bring up a unix shell window with gcc and emacs and all that is very nice, as well.
On the minus side, the machine is not a speed demon no matter what the apologists say. While sure, not many apps are optimized for altivec, but then again how many apps are optimized for MMX and SSE(n)? For most of my modest home needs, however, it has more than enough power. However, the OS doesn't feel as "real time" as it should--often clicking on something will bring up the "spinning wheel" cursor for several seconds or occasionally much more before a menu pops up. Probably adding more RAM (I have 256 currently) would address this, but it feels like something in their windowing system is blocking that shouldn't be. Also some lack of polish such as the cursor sometimes showing the wrong thing (e.g., I-beam when over a button), Finder windows occasionally defaulting to messed-up column widths, etc, that have been there through several updates is a little sketchy given that they control the entire hardware platform... another issue is that a lot of open-source stuff seems to be hard to get to build under OSX/Darwin, although this will probably improve over time. You can definitely feel that they have a lot fewer programmers than MS, but this seems to cut both ways.
I sort of resented having to pay to upgrade to 10.2 to get decent audio support, and there are still apps that haven't been ported to OS-X natively (and running classic mode feels much less seamless than running DOS and Win16 apps did under Win95) Microsoft at least made their fast-changing multimedia layer (including DirectX) seperate and freely upgradable independent of the OS, so you could run a lot of new stuff on '95 for quite a few years--with Mac it seems like you have to commit to upgrading every year or so to be able to run most programs. Also, the nice built-in apps make (brainwash?) you want everything else on your system to work/look the same (I didn't have the same problem with Office for Windows for some strange reason... ;)
I don't feel like I'm getting the absolute most "bang for my buck", but I've gotten more and more sick of fscking around with malfunctioning computers, especially in precious free time, that I'm willing to pay a premium for a pretty and reliable system. However, I still like my Win2K box at work just fine as well. I think switching between OS's every so often is a good idea for anybody just to keep from getting to attached to one paradigm...
Why not just get an optical PC-CARD or USB adapter and go digital out straight into your receiver? Those are pretty cheap, and since it's digital there theoretically shouldn't be any loss. Since you're thinking of springing for an HDTV plasma display, I'm assuming your receiver can do Dolby Digital or DTS 5.1 already and should have decent D-A converter on it.
I think a console like the PS2 is a much better place for exhibiting the old-fashioned "clever hack", because its weird and exposed architecture allows for many, many more ways to do things, some of which can be very "crazy". Plus hand-coding assembly to fit in 4K or 16K for the vector units seems very similar to the old demo contests... is there a demo scene for the PS2 linux kit? There should be if there isn't...
I like the fact that NT's 64-bit time counts something like 100s of microseconds since 1600 AD instead of the old 32-bit DOS-style "seconds since 1970". This has the nice property that it's already larger than 4 billion, so it makes it harder to screw up an implementation without seeing an error immediately. Perhaps if they made TAI not even close to UTC then it would be easier to spot bugs. Perhaps you could even just store both values in timestamps, and use TAI for computation/timestamping and UTC for display... that way you just depend on the OS to keep both values up to date, and less room for application bugs...
However, I don't think it would be possible to get a generic OpenGL implementation to run well on a PS2, even with inside information--with its 300 MHz CPU and no L2 cache, single-texturing, 4MB VRAM, no hardware clipping, etc, it would perform very poorly on platform-agnostic code and data. It would also be very slow to convert 32-bit textures into palettized on the fly unless you did a really crappy job of it. The PS2 will perform poorly unless you carefully structure your game around it's unique needs, and memory will be very scarce if your driver has to copy and translate your geometry and textures into a hardware-friendly format. PC-based graphics cards are designed around running DirectX and OpenGL efficiently--the PS2 has no such requirement. You might have a chance for some performance at the expense of memory if the game drew huge portions of the world into a display list, allowing the driver to convert this into a big DMA chain.
You've all seen how the first PS2 games often looked worse than Dreamcast games, and these were from professional developers who were coding for the PS2!
I suppose if you just wanted to run Minesweeper or something you'd be okay...
Anyways, I myself went for camber plates instead of bending my struts, although the guy I know who went the bending route has had tons of track time on that car with no problems. Since at the present time I just do open lapping sessions with no racing, I'm much more interested in mods that make the handling more neutral, keep the brakes from fading, or make the car more reliable under track conditions, rather than eaking out a little more horsepower or shaving off a few pounds.
It's interesting that the performance-oriented wing of the Japanese import crowd seems much more focused on drag racing, while their counterparts in the Bimmer/Porsche groups go for road racing as the yardstick by which to judge a sports car. Different strokes for different folks, I guess.
Actually, slightly lowering your car by cutting off a half-coil is recommended by some respected motorsports books because it not only reduces the center of gravity but slightly increases the effective spring-rate as well. I also know people who've bent their front struts to gain some negative camber, but that seems a little sketchy to me...
I agree, though, that most of these modded cars are slower than stock, although if I see one with big floppy drag tires on the front wheels I have to allow for the possibility of it being fast. My other favorite is the cars whose main modification seems to be sticking the wheels out an extra six inches or so--the thought of all that shearing force on the bearings makes me cringe...