So really, to upgrade my CPU, I need to get a new motherboard. To get a new motherboard, I probably need to get a new case & power supply, maybe some new RAM... and hell, at that point I might as well get a new computer and plug in some of my old peripherals.
You shouldn't need to replace your case and power supply, unless you have an old AT case (ATX is now standard). As for complaining about buying a new motherboard and RAM...well, it'd be stupid to put a fast, new CPU in a machine with 66MHz RAM, so really, you just keep the motherboard, CPU, and RAM together. You don't have to upgrade your video card, hard disks or monitor (all similarly expensive components) if you don't want to.
Either way I'm out $500-1000... think I'll just stick with my Celeron 366, it functions well enough...
If it does function for what you do with it, fine, keep it. But the high end is lead by early-adopters who buy hardware so they can run games, and an old PC won't cut it for those. You're not in the market this is targeted at.
He should seek professional help. Soon. That's right up there with self-mutilation.
The first thought that jumped to my mind was "can I add Generic Pre-Processor to the list of languages this works in?". I'm drawn to this task like a fly into a zapper. Please, someone write a +5 insightful essay about why I shouldn't do this...before it's too late.
The eMac has a 17 inch display. Who the hell is going to run that thing at 1280x960!?! I'm sure your eyes would be just fine running at the recommended resolution of 1024x768 @89 Hz.
You're right; 1280x960 is too skimpy. I used to use a FIFTEEN inch moniter at 1280x1024; then I used a 17" at 1280x1024 for awhile. I now use a 19" monitor at 1600x1200 (75Hz, by the way), all with standard-size fonts. Just because your eyes can't handle small pixels doesn't mean other people's can't.
I'm not entirely sure, but since Parallax developed Descent 1 & 2 for Interplay, and Volition developed the Freespace games for Interplay, I think that Interplay would be the one to get credit, at the very least for convincing the other two to release the source. Interplay is earning major karma points...
No, the credit goes to Parallax, of which Volition is one of two divisions. After Descent 2, when Parallax opened a second office, they decided to that the two offices would remain the same company, but would work on separate projects; they're known as Outrage and Volition. Volition works on Freespace, and Outrage on Descent.
Of course, Interplay may have a part, but the fact that both games were written for Interplay isn't evidence in itself.
Decent 2D games, yes. Decent 3D games, no. There is a 3D engine for the TI-89/92 called FAT (fixed-angle texturing), but to make a worthwhile shooter requires quite a bit more. Specifically, it requires memory. The author of FAT did some calculations and decided that sprites would have to be left in storage, not loaded into memory, because of size. That means they can't be compressed, which means that they take up a huge amount of space.
Also, don't overestimate the processing power on those calculators. Just because it can do multivariable calculus and run a simple 3D engine doesn't mean it can run a real 3D game; the 3D engine only works because it's using a very primitive rendering method, similar to the one used in the old Castle Wolfenstein.
Finally, I don't think there will be any commercial game development for TI calculators for political reasons. TI wants their product to be a calculator, not a portable computer. Commercial games would hurt the TI-89's credibility in its core market. Many teachers already outlaw the TI-89 because it can do things that they'd rather the students do; commercial games would just give them another reason.
One of the nice things about web pages is that you can just look at the source. Wouldn't it be great if you could look at the source for any widget in an application in the same way? Even make changes just by editing the source directly, if you wanted to. So, for instance, if I'm using a Word processor and there's a function I never use I can just delete it from the source script.
That's exactly what's so wonderful about the GPL. Of course, it's not quite so simple (you have to recompile, you have to set some things up), but the idea is the same. The difference with JS is that the source is open *by default* (since it's interpreted, and you have to run an obfuscator to close it), so many people open-source their scripts (though not in a legal sense, or in the Stallman sense) without realizing it, whereas with compiled languages like C it's closed by default (and must be opened by adding extra files to the archive).
Reasonable people agree that the creator of a work should compensated for his efforts, hence copyright - but it has no basis in the constitution.
Who modded this up? The Constitution has a clause explicitly allowing copyrights/patents. The real question is in what ways the first amendment contradicts (and therefore overrides, since amendments have precedence over body) that clause.
For example, I have the right to criticize a work, since that is speech. I have the right to pull examples (quotations) from that work to support my argument, which is also protected speech. I do not have the right to copy the work verbatim, since that is someone else's speech, not mine.
As usual, the extremes on either side of the argument need to be tempered to find a workable solution. And it isn't going to be found in the Constitution.
The corporatist extreme - restrict everything - contradicts the first amendment. That means that the only the Slashdot extreme or a moderate position can be legal.
Your post is definitely a troll, but since you immediately admit and justify it ("Since there's hardly anything to talk about with regard to the review, let's get to it!") I will respond anyways.
I'll start my argument with an anecdote. There is one particular C++ program I was writing, not atypical, and not a particularly difficult problem computationally. I was programming on my AMD Athlon 1400MHz, a fast computer by any reasonable standards, and found that it wasn't running as fast as I'd like. I profiled it, and found the functions to blame were operator overloads (called very frequently) on a class coord, which looked something like this:
friend coord operator+ (coord a, coord b) { return coord(a.x+b.x, a.y+b.y); } Spot the error on that line. Now, ask someone who's never programmed in C before to spot the error. Give up? It's the passing by value, which prevents the compiler from inlining the function; the correct way to write that function is:
friend coord& operator+ (const coord& a, const coord& b) { return coord(a.x+b.x, a.y+b.y); } If not for my experience programming in C, I never would've realized that.
It is my observation that people who are taught to ignore C, and start immediately with an object-oriented language such as C++, <B>never</B> learn low-level concepts such as inlining or pointers, and never learn to truly understand what it is that they're writing. I say that the best way to learn and properly appreciate the "right way" of doing something is to first do it the *wrong way*, in a project that doesn't matter for anything. I have debugged pointer spaghetti, written code with dozens of meaninglessly named global variables with no comments, written procedural code, and had I not done these things, I wouldn't know why it is important to name variables intelligently, to use object orientation, or to use pointers carefully. You tell someone to comment their code, and you get lines line this: a=b;// Copy b into a The only reason the comment is there is because they were taught to always comment their code, to comment every line, etc. On the other hand, someone who's dealt with uncommented code before would put useful comments where they need to be.
I agree with you that C should not be used in production code where it can be avoided (that is, areas other than systems and embedded programming). However, I strongly believe that people should always learn and master C before learning higher-level languages. If the only reason you use classes is because you were taught that that's how to write clean code, then you're not using them correctly. On the other hand, if you're using classes because you wrote in C up to the point where you encountered a problem that required inheritance or polymorphism, then you're using the feature for the right reasons.
Your example, by the way, in which a loop that increments elements of an array is parralelized in hardware, is actually simpler than you think. The compiler first performs loop unrolling (a very, very old idea), then analyzes the code blocks to see that they don't work with the same data, and parralelizes them. Your particular example implies that the only real way to solve parralelism is to define parralel functions and human-solve them; this clearly violates the distinction between language and library, and doesn't really help. Besides, in your case example of compiling high-level code into hardware, I could come up with far more examples where object orientation hurts rather than helps. OO hides all of the overhead, promoting huge bloat which, while not a problem in software, is fatal in hardware.
You completely missed the point of that post; the entire point of flags is to modify a pointer, as meta-data, not to point into 16 exabytes of non-existant memory, which would be useless. Where having a few extra bits for each pointer is useful is when doing things like reference counting (wouldn't it be interesting to have that in hardware? Set a bit to turn it off), caching (reserve some bits to indicate which caches have copies), memory management schemes (compiler hints, memory handles, virtual memory info, etc), and so on. Don't assume that just because every value of a 32-bit pointer corresponds to a unique memory location that the same must by true of a long long pointer.
I don't wan't to be another BG, but I think this time he's over-cautious. Filling 2^64 bytes of memory, over a 66MHz/64 bit bus would take about 132 billion (10^9) seconds
That's true, but only relevant if you assume that 64-bit addressing gets you 2^64 bytes of memory. In reality, many of those bits are used for flags and other non-addressing things, so assume more like 2^48 bytes of memory. Assume a 266MHz/64 bit bus (4x your example, more like what's presently available on consumer-level machines), and that's 36.7h to initialize. Considering specifically the case of large mainframes with many processors, where this will first become an issue, divide by 128 processors/memory banks, with each of those going at full speed only 17.2m is necessary - and that's without even considering how much faster and wider buses will be by the end of the decade. I think you're too optimistic; assumptions like those are what got us narrow address spaces in the first place.
Well in that case, what about Kali, or similar programs? That doesn't have any copy protection, and people certainly can play there to get around CD keys. Why wouldn't Blizzard just require bnetd to do authentication (through Blizzard's servers)?
This action makes no sense. Battle.Net is not a money-maker for Blizzard; if anything, it is an expense, despite the presence of banner ads, made to boost sales of the games. Blizzard was dependent on third-party matchmaking utilities from the beginning, with Kali (a pay service, no less!), and Blizzard supported them (by including a copy of Kali on the CDs); had there been no third-party applications for multiplayer Warcraft, there would've been no multiplayer Warcraft at all, and much less of a community for Starcraft to take hold in. So, what's their reason for shutting down bnetd? Unrestrained lawyers are the only reasonable explanation; it is clearly to Blizzard's financial disadvantage. I expect that Blizzard will back off when higher-ups find out about this.
I noticed that in the article, the author mentions Moore's Law as stating that transistor densities double every 18-24 months. Wasn't it originally 12 months, then changed to every 12-18 months?
Re:Currently Xbox emulation is infeasible.
on
X-Box Emulated (Not)
·
· Score: 4, Interesting
Even if you could recompile whole blocks of code into tightly optimized emulation loops, there are many more things that need consideration on the emulation side such as IO ports, HD access, interrupts, proprietary 3D hardware - so what if DirectX is used? Its interface to the actual graphics chip does not exist on the PC side, so you'd have to emulate all that as well.
IO ports and HD access shouldn't pose a problem because they're not speed critical (HD would actually be faster, because of faster drives), and the 3D hardware is not the problem you seem to think it is. The "actual graphics chip" *does* exist on the PC side; nVidia sells near-identical cores to Microsoft and to PC motherboard manufacturers. Plus, the games are probably only interfacing with a DirectX library implementation somewhere on the XBox hard disk, so that can be trapped and redirected into the PC DirectX library.
As for the N64, the UltraHLE emulator runs a large percentage of games; yes, it does so with some hacks, but on consoles, the only programs you need consider are the successful commercial ones, not thousands of freeware programs with thousands of different sets of quirks.
This is also why the high level emulators will never run more than Mario 64 and Zelda 64 without ugly hacks, since both the CPU, graphics and sound chips can be reprogrammed and none of the current emulators can handle this.
DirectX HAL comes to the rescue here; the games shouldn't be touching the hardware except through DirectX, to which calls can be trapped.
I never said emulating Xbox will always remain impossible. At this time however, because of current CPU speeds and the sheer complexity of the Xbox system, you cannot expect to see an emulator. Not for at least five years, probably closer to ten.
I've gotten the distinct impression that Microsoft started with a PC and modified it until they had something suitable to sell as a console. Yes, it will certainly take time to reverse-engineer the thing to the point where it can be emulated. However, Microsoft's laziness may well mean that the XBox and PCs are surprisingly similar. I also think that the hardware of five to ten years from now will be absolute overkill for emulating an XBox.
Of course, until someone cracks the XBox BIOS, we're both on speculation, which makes this argument rather pointless.
It's very interesting how you came to that $2K/month number. In fact, looking over the components of that, I object to *every* number you put into that.
I'd expect 24/7 pay-per-view access
$10 and up per pay-per-view item... probably $200-$300 worth of use.
He didn't say "24/7 FREE pay-per-view access", and neither did he say he would be using it 40-60 hours/month, as your number implies.
24/7 porno
A few nights a week at $8 per movie - another $100 or more.
Where'd that $8/movie number come from? And how is this hypothetical buyer possibly going to have free time for this *and* 50 hours per month of pay-per-view?
any On-Demand movie I want for free
Another $100 or more...
For $100/month, you could go to a theater every time you felt like watching a movie, and watch it on much better equipment. Off by a factor of 3 or 4.
and every single channel they can cram in the cable band.
Licensing and fees to the subscription channel providers = perhaps $200 or more depending on your market.
Off by a factor of five or more. Current cable providers fill the cable as it is already; you don't see them getting away with $200/month, do you?
I also expect an unrestricted U/L and D/L line on my Internet connection
UUNET/Sprint T1 = $800/month...
Everyone knows that T1 prices are a joke. Also, the price you're quoting includes business-class service (which you added MORE cost on for later), non-trivial installation, and 100% bandwidth use (which no home user reaches), and is rediculous anyways. Compare against business-class uncapped DSL to be more reasonable.
the ability to put up a server
See above (included)
"Ability to put up a server" and "ability to put up a 5-million-hit-a-day web server" are completely different things. Many DSL providers give you this privelege, so long as you don't abuse it, and you certainly don't need a T1 for it.
tech support for any computer problem I have
Reasonable rate of $65/hour, assuming you're calling only during office hours. Reasonable estimate of 5 hours/month = around $250...
The only reason for 5 hours/month is if either (a) 4 of them are on hold, (b) the support is extremely incompetent, or (c) the service gives you too many problems in need of supporting. Either way, that's completely unacceptable for $65/hour, so your "reasonable estimate" of 5 hours/month is completely unreasonable.
and a 99.9% guaranteed uptime on the line
SLA for UUNET/Sprint. See above. Definitely business grade T1 service.
Three-nines uptime is "business grade T1 service"? Add one more nine to that, maybe two for a higher price. 43 minutes downtime per month would be consumer-level standard if the DSL providers weren't so blatantly incompetent.
And I want caller ID, call waiting, every single other feature on my phone, the ability to block business (telemarketer) calls, and the best voicemail system known to man.
At least another $100.
Actually, this is more like $0/month, plus a one-time bill for a fancy phone with an LCD. None of these actually cost the provider any substantial amount, and they're certainly not worth $100/month.
TOTAL BILL: $2,000+ / month
You added another $250 in rounding - and, of course, all the numbers you used to get there were bogus anyways.
And you want this for $200? What the hell are you paying with, Flooz? You'll probably have similar results...
$200 is low, but it's in the right ballpark. Your figure is rediculous; why it's at +5 is beyond me.
At the moment, no computer on this planet has enough juice to emulate the Xbox (No, not even the supercomputers which have 9,600 CPUs - because multiple CPUs don't make it any faster to emulate a single CPU)
That is absolutely false. The XBox's CPU is just an Intel CPU, which most home computers have something similar to, so not much needs to be done to emulate that (maybe emulate a few instructions, or shift opcodes). The video is handled through DirectX and an nVidia video chip; again, most people already have something similar, and minimal translation is necessary.
People said that emulating the Nintendo 64 was impossible, but that was done, not by emulating the hardware at a low level, but through high-level emulation. Compatibility is slightly less, but it's orders faster.
The fact that the XBox hasn't been fully reverse-engineered yet is an obstacle to making an emulator, yes, and this particular "emulator" is clearly a hoax, but it is by no means impossible.
The site's not responding for me (Slashdotted? big site for that), so I'm going by the summary, which *completely* misses the mark with broadband's failures. Broadband in the U.S. is failing for two reasons: the infrastructure is owned by companies who are neither competent to nor motivated to provide broadband, and population densities are such that updating antiquated infrastructure is expensive.
Consider the telcos, who are responsible for providing DSL. They want DSL dead, because it cuts into their massive-profit sales of T1s. They're also big, lumbering bureaucracies, which deal badly with change. I won't recount my own DSL horror stories, but there are plenty to be had at DSL Reports. Technically DSL is functional and capable, but the businesses behind it, and the support bureaucracies, are not.
Cable has different problems. First, there's the cable companies; in my area, and in others, cable Internet is simply not an option because the local providers don't offer it. There's also the problem of bandwidth sharing. It's true that DSL bandwidth is also shared, but it's shared at a central point, which is easily upgraded; with cable, mis-estimation of demand or usage can leave people drastically short on bandwidth. (DSLReports again for horror stories).
Finally, consider the population layout in the US, as compared to elsewhere. If you have population-dense cities, surrounded by low-density farmland, you can provide access to most of the population simply by providing short-range access in the cities. In the US, most of the demand is in the suburbs, which involve much longer distances and are, therefore, much harder to provide for. (This is especially true in my home state of Massachusetts, where economics are such that the demand and the money is all in the suburbs).
Further, a tag would give governments and law enforcement agencies a means to literally "follow the money" in illegal transactions. (from article)
Anyone else disturbed by this? Previously, while credit cards, banks, checking, and money transfers involve giving up privacy with your purchases, cash was an anonymous, almost universally accepted form of payment. What's to stop a retailer from reading the tags on the bills they get to see who their customers are, and spam them? What about banks, where all currency eventually ends up? There's a lot of potential to use this for tracking people's purchases, and that's a bad thing.
You must be joking...this is one of the worst strategies I've heard in awhile. It's not even clever or funny, like the science vessel rush. Do you realize how expensive archons are, in vespene? And that they have no range? That they get taken down in one shot by an EMP pulse, are outranged by almost everything (tanks, carriers, hydralisks, guardians), cost a fortune, and don't arrive until the end of the tech tree? Or, were you talking about the joke of a campaign AI?
Don't tell me that Chess has no strategy because you found a trick which some silly AI from 1990 will consistently fall for. Go and play some skilled people, and be taught a painful lesson. My hydralisks are looking forward to it.
The last time I played with someone actually used a strategy besides simply building a lot of medium units and some large units and then sent them all as soon as possible was.. well, never.
Then you need to find a better group of opponents. Just as it is possible to get an army together and send it off with no strategy in real war, it's possible to do that in game war. It just isn't very effective. In the case of the game StarCraft, it's critical that you get the right mix of units together and use them properly. Suppose, for example, I built a fleet of battlecruisers and sent them in my enemy's direction. Whoops, a group of hydralisks and devourers destroyed them all while I wasn't looking. Suppose instead I send those battlecruisers around the back, into my opponent's mineral patch. I could do a huge amount of damage with those cruisers.
Practicing good strategy and tactics isn't technically necessary, but someone who makes major strategic errors loses games. Sure, it would be nice if there was a model for supply lines and moving supplies around. I bet Napoleon thought the same thing in Russia.
Star Trek falls into cliches. It kills off one-appearance characters practically every episode while leaving the major characters alone. It makes up new science whenever it needs a plot device. I personally feel that Trek has gone down hill since Voyager, and I'm not terribly optimistic about the new series. However, Star Trek is not guilty of the complete disregard for science you accuse it of. When possible, the writers do come up with plausible explanations.
Spock's pure logic: This is literally impossible. Biological brains are based on pattern matching, which necessitates illogical responses.
Actually, Vulcans are naturally more emotional and illogical than humans are, but suppress illogic and emotion through meditation and training. A human could learn to do this, too, especially given a 200-year lifespan to do it.
The dilithium crystals: As a fuel source, these are contrived beyond belief. Any good crystallographer knows that crytalline structures are too inert to supply decent energy returns. Try burning a diamond, if you don't believe me.
The dilithium crystal is not a fuel source, it regulates the reactions which power the ship. The real fuel source is deuterium and anti-matter; reactions of these do release a lot of energy.
Warp factor 9: The idea that they could exceed the speed of light exactly nine times is ludicrous. As you move further from the speed of light, the rate at which speed increases grows immeasurably larger. It would be impossible to achieve any reasonable system of measurement at these speeds.
The warp scale is not linear. Warp 1 is the speed of light. Warp 10 is infinite speed; Warp factors approaching warp 10 approach infinite speed.
No plants on the enterprise: Anyone else notice this? You need plants to breathe, fools.
When you're light years away from any star, providing the light to keep plants alive gets non-trivial. And, do you seriously believe that photosynthesis is the *only* reaction which can convert CO2 into O2?
Artificial gravity: This was never explained. In any series.
True, but just about every sci-fi series has to have it, because it's simply not practical to film in zero gravity.
It's not entirely clear whether they intend to do this only for university-owned computers or for student-owned computers as well. If this is only for the school's computers, I don't see too much problem with it. (I would see a problem if it turned into anything more than a screensaver, though). After all, it's the school's hardware, and a screensaver really can't interfere with work.
If they're talking about putting it on machines that belong to students, then this is objectionable in the extreme. Students have the right to control what software runs on the hardware they pay for, and I can imagine bad things happening when faculty demand to install it on incompatible platforms such as Linux.
Hence if you've got a 500Mhz chip with 2 stages and the clock physically placed near stage 1, then stage 1 of the pipeline will run at 500MHz, stage 2 will also run at 500MHz but with some latency, so the two-stage pipeline will complete an instruction very slightly over 2 cycles. Add more stages, you'll get a bigger effect at the end. And as clock speeds go faster, you'll eventually hit the ceiling -- the latency might actually be as fast as a single cycle itself.
Latency doesn't affect the time required to complete the instruction, only the time at which it is executed. If the clock reaches second pipeline stage late, the time required to complete that pipeline step is latency+calculation time. If that's more than one cycle, the chip is clocked higher than it can run, period. Anyways, there's a fairly obvious solution to clock latency in pipelines. Put the start of the pipeline near the registers, and the end of the pipeline near the registers, then a U shape for all the ones in the middle; since a stage only needs to interact with the ones before and after it, which are physically adjacent, the effective latency is small.
[from article] But after a point, cranking up the clock speed becomes an exercise in diminishing returns. That's why a one-gigahertz chip doesn't run twice as fast as a 500-megahertz chip.
Wrong. A 1GHz chip doesn't run twice as fast as a 500MHz chip because of pipelining, and because the support infrastructure in a typical PC can't handle a 1GHz chip well, so it spends a lot of time waiting for hard disk access and memory. Eliminating the clock isn't going to make the heads on a hard disk move any faster. The real benefit is that an idle component can't be a bottleneck anymore.
This is waaay too late for anyone but the parent poster to see it, but I just have to say it.
The N64 controller is, IMHO, the most disgusting, un-usable and painful controller I've ever used. Basically, it makes two assumptions which completely fail to fit me: short finger length (the joystick sits at the bend of my thumb, where it doesn't belong and isn't usable), and right-handedness (offer a version with the whole thing mirrored and it'd be better). At the same time, a large part of the controller goes unused and seems to serve no purpose other than balance (the D-pad and left shoulder button). I played the N64's games emulated on a PC, with a joystick (one-handed, much more ergonomic, and plenty of buttons). I'm really not a console person, but even if I was, an N64-like controller would be more than enough to drive me away from Cube. What consoles really should have, is a standardized (every console seems to make its own) and interchangable interface (USB?) for controllers.
Slashdot Mirror
Of course, Interplay may have a part, but the fact that both games were written for Interplay isn't evidence in itself.
Decent 2D games, yes. Decent 3D games, no. There is a 3D engine for the TI-89/92 called FAT (fixed-angle texturing), but to make a worthwhile shooter requires quite a bit more. Specifically, it requires memory. The author of FAT did some calculations and decided that sprites would have to be left in storage, not loaded into memory, because of size. That means they can't be compressed, which means that they take up a huge amount of space.
Also, don't overestimate the processing power on those calculators. Just because it can do multivariable calculus and run a simple 3D engine doesn't mean it can run a real 3D game; the 3D engine only works because it's using a very primitive rendering method, similar to the one used in the old Castle Wolfenstein.
Finally, I don't think there will be any commercial game development for TI calculators for political reasons. TI wants their product to be a calculator, not a portable computer. Commercial games would hurt the TI-89's credibility in its core market. Many teachers already outlaw the TI-89 because it can do things that they'd rather the students do; commercial games would just give them another reason.
For example, I have the right to criticize a work, since that is speech. I have the right to pull examples (quotations) from that work to support my argument, which is also protected speech. I do not have the right to copy the work verbatim, since that is someone else's speech, not mine.
The corporatist extreme - restrict everything - contradicts the first amendment. That means that the only the Slashdot extreme or a moderate position can be legal.
Your post is definitely a troll, but since you immediately admit and justify it ("Since there's hardly anything to talk about with regard to the review, let's get to it!") I will respond anyways.
// Copy b into a
I'll start my argument with an anecdote. There is one particular C++ program I was writing, not atypical, and not a particularly difficult problem computationally. I was programming on my AMD Athlon 1400MHz, a fast computer by any reasonable standards, and found that it wasn't running as fast as I'd like. I profiled it, and found the functions to blame were operator overloads (called very frequently) on a class coord, which looked something like this:
friend coord operator+ (coord a, coord b) { return coord(a.x+b.x, a.y+b.y); }
Spot the error on that line. Now, ask someone who's never programmed in C before to spot the error. Give up? It's the passing by value, which prevents the compiler from inlining the function; the correct way to write that function is:
friend coord& operator+ (const coord& a, const coord& b) { return coord(a.x+b.x, a.y+b.y); }
If not for my experience programming in C, I never would've realized that.
It is my observation that people who are taught to ignore C, and start immediately with an object-oriented language such as C++, <B>never</B> learn low-level concepts such as inlining or pointers, and never learn to truly understand what it is that they're writing. I say that the best way to learn and properly appreciate the "right way" of doing something is to first do it the *wrong way*, in a project that doesn't matter for anything. I have debugged pointer spaghetti, written code with dozens of meaninglessly named global variables with no comments, written procedural code, and had I not done these things, I wouldn't know why it is important to name variables intelligently, to use object orientation, or to use pointers carefully. You tell someone to comment their code, and you get lines line this:
a=b;
The only reason the comment is there is because they were taught to always comment their code, to comment every line, etc. On the other hand, someone who's dealt with uncommented code before would put useful comments where they need to be.
I agree with you that C should not be used in production code where it can be avoided (that is, areas other than systems and embedded programming). However, I strongly believe that people should always learn and master C before learning higher-level languages. If the only reason you use classes is because you were taught that that's how to write clean code, then you're not using them correctly. On the other hand, if you're using classes because you wrote in C up to the point where you encountered a problem that required inheritance or polymorphism, then you're using the feature for the right reasons.
Your example, by the way, in which a loop that increments elements of an array is parralelized in hardware, is actually simpler than you think. The compiler first performs loop unrolling (a very, very old idea), then analyzes the code blocks to see that they don't work with the same data, and parralelizes them. Your particular example implies that the only real way to solve parralelism is to define parralel functions and human-solve them; this clearly violates the distinction between language and library, and doesn't really help. Besides, in your case example of compiling high-level code into hardware, I could come up with far more examples where object orientation hurts rather than helps. OO hides all of the overhead, promoting huge bloat which, while not a problem in software, is fatal in hardware.
You completely missed the point of that post; the entire point of flags is to modify a pointer, as meta-data, not to point into 16 exabytes of non-existant memory, which would be useless. Where having a few extra bits for each pointer is useful is when doing things like reference counting (wouldn't it be interesting to have that in hardware? Set a bit to turn it off), caching (reserve some bits to indicate which caches have copies), memory management schemes (compiler hints, memory handles, virtual memory info, etc), and so on. Don't assume that just because every value of a 32-bit pointer corresponds to a unique memory location that the same must by true of a long long pointer.
Well in that case, what about Kali, or similar programs? That doesn't have any copy protection, and people certainly can play there to get around CD keys. Why wouldn't Blizzard just require bnetd to do authentication (through Blizzard's servers)?
This action makes no sense. Battle.Net is not a money-maker for Blizzard; if anything, it is an expense, despite the presence of banner ads, made to boost sales of the games. Blizzard was dependent on third-party matchmaking utilities from the beginning, with Kali (a pay service, no less!), and Blizzard supported them (by including a copy of Kali on the CDs); had there been no third-party applications for multiplayer Warcraft, there would've been no multiplayer Warcraft at all, and much less of a community for Starcraft to take hold in. So, what's their reason for shutting down bnetd? Unrestrained lawyers are the only reasonable explanation; it is clearly to Blizzard's financial disadvantage. I expect that Blizzard will back off when higher-ups find out about this.
I noticed that in the article, the author mentions Moore's Law as stating that transistor densities double every 18-24 months. Wasn't it originally 12 months, then changed to every 12-18 months?
As for the N64, the UltraHLE emulator runs a large percentage of games; yes, it does so with some hacks, but on consoles, the only programs you need consider are the successful commercial ones, not thousands of freeware programs with thousands of different sets of quirks. DirectX HAL comes to the rescue here; the games shouldn't be touching the hardware except through DirectX, to which calls can be trapped. I've gotten the distinct impression that Microsoft started with a PC and modified it until they had something suitable to sell as a console. Yes, it will certainly take time to reverse-engineer the thing to the point where it can be emulated. However, Microsoft's laziness may well mean that the XBox and PCs are surprisingly similar. I also think that the hardware of five to ten years from now will be absolute overkill for emulating an XBox.
Of course, until someone cracks the XBox BIOS, we're both on speculation, which makes this argument rather pointless.
He didn't say "24/7 FREE pay-per-view access", and neither did he say he would be using it 40-60 hours/month, as your number implies.
Where'd that $8/movie number come from? And how is this hypothetical buyer possibly going to have free time for this *and* 50 hours per month of pay-per-view?
For $100/month, you could go to a theater every time you felt like watching a movie, and watch it on much better equipment. Off by a factor of 3 or 4.
Off by a factor of five or more. Current cable providers fill the cable as it is already; you don't see them getting away with $200/month, do you?
Everyone knows that T1 prices are a joke. Also, the price you're quoting includes business-class service (which you added MORE cost on for later), non-trivial installation, and 100% bandwidth use (which no home user reaches), and is rediculous anyways. Compare against business-class uncapped DSL to be more reasonable.
"Ability to put up a server" and "ability to put up a 5-million-hit-a-day web server" are completely different things. Many DSL providers give you this privelege, so long as you don't abuse it, and you certainly don't need a T1 for it.
The only reason for 5 hours/month is if either (a) 4 of them are on hold, (b) the support is extremely incompetent, or (c) the service gives you too many problems in need of supporting. Either way, that's completely unacceptable for $65/hour, so your "reasonable estimate" of 5 hours/month is completely unreasonable.
Three-nines uptime is "business grade T1 service"? Add one more nine to that, maybe two for a higher price. 43 minutes downtime per month would be consumer-level standard if the DSL providers weren't so blatantly incompetent.
Actually, this is more like $0/month, plus a one-time bill for a fancy phone with an LCD. None of these actually cost the provider any substantial amount, and they're certainly not worth $100/month.
You added another $250 in rounding - and, of course, all the numbers you used to get there were bogus anyways.
$200 is low, but it's in the right ballpark. Your figure is rediculous; why it's at +5 is beyond me.
That is absolutely false. The XBox's CPU is just an Intel CPU, which most home computers have something similar to, so not much needs to be done to emulate that (maybe emulate a few instructions, or shift opcodes). The video is handled through DirectX and an nVidia video chip; again, most people already have something similar, and minimal translation is necessary.
People said that emulating the Nintendo 64 was impossible, but that was done, not by emulating the hardware at a low level, but through high-level emulation. Compatibility is slightly less, but it's orders faster.
The fact that the XBox hasn't been fully reverse-engineered yet is an obstacle to making an emulator, yes, and this particular "emulator" is clearly a hoax, but it is by no means impossible.
The site's not responding for me (Slashdotted? big site for that), so I'm going by the summary, which *completely* misses the mark with broadband's failures. Broadband in the U.S. is failing for two reasons: the infrastructure is owned by companies who are neither competent to nor motivated to provide broadband, and population densities are such that updating antiquated infrastructure is expensive.
Consider the telcos, who are responsible for providing DSL. They want DSL dead, because it cuts into their massive-profit sales of T1s. They're also big, lumbering bureaucracies, which deal badly with change. I won't recount my own DSL horror stories, but there are plenty to be had at DSL Reports. Technically DSL is functional and capable, but the businesses behind it, and the support bureaucracies, are not.
Cable has different problems. First, there's the cable companies; in my area, and in others, cable Internet is simply not an option because the local providers don't offer it. There's also the problem of bandwidth sharing. It's true that DSL bandwidth is also shared, but it's shared at a central point, which is easily upgraded; with cable, mis-estimation of demand or usage can leave people drastically short on bandwidth. (DSLReports again for horror stories).
Finally, consider the population layout in the US, as compared to elsewhere. If you have population-dense cities, surrounded by low-density farmland, you can provide access to most of the population simply by providing short-range access in the cities. In the US, most of the demand is in the suburbs, which involve much longer distances and are, therefore, much harder to provide for. (This is especially true in my home state of Massachusetts, where economics are such that the demand and the money is all in the suburbs).
You must be joking...this is one of the worst strategies I've heard in awhile. It's not even clever or funny, like the science vessel rush. Do you realize how expensive archons are, in vespene? And that they have no range? That they get taken down in one shot by an EMP pulse, are outranged by almost everything (tanks, carriers, hydralisks, guardians), cost a fortune, and don't arrive until the end of the tech tree? Or, were you talking about the joke of a campaign AI? Don't tell me that Chess has no strategy because you found a trick which some silly AI from 1990 will consistently fall for. Go and play some skilled people, and be taught a painful lesson. My hydralisks are looking forward to it.
Practicing good strategy and tactics isn't technically necessary, but someone who makes major strategic errors loses games. Sure, it would be nice if there was a model for supply lines and moving supplies around. I bet Napoleon thought the same thing in Russia.
Actually, Vulcans are naturally more emotional and illogical than humans are, but suppress illogic and emotion through meditation and training. A human could learn to do this, too, especially given a 200-year lifespan to do it.
The dilithium crystal is not a fuel source, it regulates the reactions which power the ship. The real fuel source is deuterium and anti-matter; reactions of these do release a lot of energy.
The warp scale is not linear. Warp 1 is the speed of light. Warp 10 is infinite speed; Warp factors approaching warp 10 approach infinite speed.
When you're light years away from any star, providing the light to keep plants alive gets non-trivial. And, do you seriously believe that photosynthesis is the *only* reaction which can convert CO2 into O2?
True, but just about every sci-fi series has to have it, because it's simply not practical to film in zero gravity.
It's not entirely clear whether they intend to do this only for university-owned computers or for student-owned computers as well. If this is only for the school's computers, I don't see too much problem with it. (I would see a problem if it turned into anything more than a screensaver, though). After all, it's the school's hardware, and a screensaver really can't interfere with work.
If they're talking about putting it on machines that belong to students, then this is objectionable in the extreme. Students have the right to control what software runs on the hardware they pay for, and I can imagine bad things happening when faculty demand to install it on incompatible platforms such as Linux.
This is waaay too late for anyone but the parent poster to see it, but I just have to say it.
The N64 controller is, IMHO, the most disgusting, un-usable and painful controller I've ever used. Basically, it makes two assumptions which completely fail to fit me: short finger length (the joystick sits at the bend of my thumb, where it doesn't belong and isn't usable), and right-handedness (offer a version with the whole thing mirrored and it'd be better). At the same time, a large part of the controller goes unused and seems to serve no purpose other than balance (the D-pad and left shoulder button). I played the N64's games emulated on a PC, with a joystick (one-handed, much more ergonomic, and plenty of buttons). I'm really not a console person, but even if I was, an N64-like controller would be more than enough to drive me away from Cube. What consoles really should have, is a standardized (every console seems to make its own) and interchangable interface (USB?) for controllers.