Yes, but how do you do the patent search? Typically the patents are phrased in such obscure but broad language that even a professional patent search is unlikely to turn everything up. Plus, paradoxically doing such a search actually increases your liability if you should overlook something. As for designing around things, that might work if so many patents on obvious things weren't granted. Inventors shouldn't have to design around things that are intuitive in their field.
Proof of concept and it doesn't really do anything, but it compiles just fine. I don't see a problem here. A real implementation would probably do some clever stuff so that the optimizer can optimize away the intermediate data structure.
You're arguing that people should user proper English on the Interne because it's for the greater good? If everyone thought like that, communism might would work. The fact is, most people don't think like that.
Nah. Nothing so esoteric as that. Just simple courtesy to your fellow man, that's all. That's how most etiquette originated: trying to save each other a bit of trouble. It worked, too, once upon a time.
The reason why 'u' and 'r' started appearing was mostly because of the speed issue.
I've always found the speed argument silly and selfish. Sure it might be slightly faster for the writer to omit letters that are technically redundant from an information theory perspective, but it cost each and every reader more time to "decode" the message. For something that's going to be read by a number of people, the aggregate time savings to everyone is net negative. I've always felt it to be truely selfish and lazy on the part of the writer and shows no consideration for the readers. If the writer doesn't have time to draft a proper message, maybe they shouldn't be posting.
But maybe I'm just a curmudgeon who'd rather see people use the extra time it takes to write properly to think through what they're saying.
Not only is it not useful but it's wrong too! The code modifies x twice between sequence points. It'll only add one to x if you're lucky, but it's behaviour is actually undefined in C/C++. "Nasal demons" and all that jazz.
1. You have multiple frames to work with. You can do inter-frame interpolation to pull more information out of stills. Ever wondered why a lower resolution is acceptable for moving pictures than stills? It's because your brain does this.
Sure. This is called superresolution, a technique pioneered by NASA for enhancing low-res images returned by probes. It works well for shots of still scenes since you compute fairly exact image registration. Doing it on fast-moving subjects such as a license plate reflected from the windshield of another moving car is baloney. It's going to be especially poor on subjects that move non-rigidly since there's no simple affine transformation.
2. You know what people, cars, etc look like. If you have n images that could have been degraded to look like a pixelated face, you can discard the m that don't look like a face at all. The n-m is then a much smaller search space, and you can make guesses within this.
If the information isn't there in the data, it simply isn't there and there's nothing you can do to bring it back. Minimizing the search space isn't going to help -- that's just throwing out more information. I've seen papers on using models of people and cars or things ling Terzopoulos' active contour models to track them spatially as they move through a scene, but that's a different task.
I believe the grand parent was referring to Philips' position of refusing to license their CD-DA logo to discs that don't conform to the Red Book standard. If the logo is absent from the disc, this suggest that it's in non-conformance, likely due to DRM. Even if the logo is present, however, that's no guarantee that it is DRM free, but the chances are better if the logo is there.
Because 3 floats/pixel adds up really quickly. You can do it now, but that was horrendously expensive back in the day. (Heck, 24bpp was considered insanely expensive once upon a time. That's why people used hacks like palettes and indexed images.)
There's also the question of how you flatten that down to the gamut of the monitor. You're typical monitor has maybe 2 to 3 orders of magnitude. Figuring out the best way to compress HDR down to that scale (i.e. tone-mapping) is still an active area of research in computer graphics even though folks here are acting like it's a solved problem.
My argument is that one cannot assume, in a multi-architecture environment, that such optimizations will always be available. Semantically, the expression array[index] must always be expanded to *(array + index) when the index is variable. In other words, the expression cannot be reduced further, because the value of the index is unknown at run time.
Yes, semantically array[index] has to have the same effect as *(array+index). But the compiler is free to generate conceptually equivalent code in any way that it pleases. Any decent C/C++ compiler optimizer that can perform strength reduction ought to be able to see how the index changes the memory location and turn it into simple pointer incrementation accordingly. And strength reduction is a well known optimization that's been around for ages -- if memory serves, even the old Red Dragon book talks about how it works in this context. If your compiler can't handle this you need to find a better compiler.
Lua at least already has something like this. They're called asymmetric coroutines. Also check out Scheme, Scheme's had continuations forever and some folks have done interesting things with them.
I subscribe to Dr. Dobbs Journal (a.k.a DDJ) and the C/C++ Users Journal (CUJ). I liked DDJ a lot when I first started subscribing to it 15 years ago. It got a little lean a few years ago but seems to be getting better lately. Still, though, it's a classic and the CD-Rom archive is terrific. CUJ is more focused on C++, but I like it for keeping up with modern trends in the C++ world.
I'd love to see a language (or language extension) cleanly define a way to let me define a code block attributes which could affect how and where it gets executed. The runtime library could then distribute that block as the environment best allows.
Have a look at OpenMP. Granted, it's more for shared-memory systems than clusters, but it works similiarly to what you describe.
Uggh... minor orthographic issues I can forgive but this one drives me nuts! Once and for all, the phrase is "for all intents and purposes." Thank you.
I have a quote that I added to my fortune file of a bit in "The Rebel Angels" by Robertson Davies where a character (the head of a huge banking firm) is explaining exactly this to his date:
"Money's something you shove around, like electricity."
"Like electricity?"
"Like large power-grids, and transformers, and that sort of thing. The diffusion of electricity is an extremely important kind of engineering. You decide where to put the energy, and how to get it there, according to the result you expect. Money is a form of power."
-- "The Rebel Angels" Robertson Davies
Yes. I have some soundtracks to games I've never played, but still find the soundtracks quite enjoyable. Of course, these are often from games in a series where I have played some of them, just not all of them. (e.g. Castlevania and Final Fantasy). So I'll admit that I'm predisposed to like them.
Indeed! I always thought that FF4 was the perfect length. Long enough to be satisfying when I beat it, but not so long that I couldn't beat it in a few weeks. FF5 and FF6 were a bit longer but still reasonable. The switch to the PS was about where they lost me.
Part of my problem is that I enjoyed the old FF games for the story. My problem is that if a game starts taking me too long to beat (not having as much free time anymore doesn't help) and tries to cram too much story in, I begin to forget the parts of the game's story before I can finish it. That's not so much fun for someone like me.
"...and the upcoming Unreal Tournament 3"
So if he is talking about Unreal Tournament 2007, that would be Unreal Tournament 4 in the numbering scheme of how many Unreal Tournaments there have been. 1 would be Unreal Tournament (1999) 2 is Unreal Tournament 2003, 3 is Unreal Tournament 2004. Finally 4 is Unreal Tournament 2007.
I suspect he meant the upcoming Unreal Engine 3. Unreal and UT99 used the Unreal Engine 1. UT2k3 and UT2k4 both used the same basic Unreal Engine 2. UT2k7 will be based on the new Unreal Engine 3. Epic's Unreal Technology Roadmap has the details.
I'd also suggest the red book as a companion/reference. The OpenGL consortium's site also has some some documentation and tutorials (e.g. the spec downloadable in PDF form.)
I'm a "gentile" who moved here to Salt Lake City at the beginning of the year to go back to grad school and this is pretty much how it's looked to me so far. About the only thing I have to add is that I've been told that many people aren't quite as wealthy as they seem at first glance here. The LDS tendency to mary young often leads to people overextending their credit trying to purchase a house too early.
And yeah, I have yet to really meet annoying pushy LDS folks like I have evangelicals. Polite protest of disinterest have worked just fine.
There also turn out to be some surprising advantages to the oddities in legislation here. For example, the recent SCOTUS Eminent Domain ruling has no effect here as the state government recently passed legislation banning eminent domain for redevelopment purposes (except, in a twisted joke, for the redevelopers own property.)
I should mention, however, that my views are somewhat coloured by a fairly centrists view of American politics and religion, so the politics and religion here don't really bother me as much as they might some. (Honestly, for the moment I find them a somewhat refreshing change from my old heavily blue state.)
Yep, that's the way my UT2k4 iCTF clan handles it. We've found most of the anti-cheat mods out there tend to make the game lag horribly and don't even catch many cheaters anyway. So we've taken to just having a large group of admins around to keep an eye on things and hand out bans.
Social problem, social solution...
(Of course, why people still try to cheat when they see players with our tag around is still a mystery to me!)
I've been looking at senior software developer positions, but is that too high up the ladder for someone 'fresh' to cope with?
Depends on where you work. I worked for several years after college and then went back to school full time for an advanced degree. At least in my experience, there's a world of difference between what the senior software engineers did and the kind of development that I do in grad school.
Most academic types don't have to worry about making their code bulletproof, "productizing" it, requirements documents, tech specs, working with UI folks, working with QA folks and bug DBs, or coding to a schedule as part of team. Then there's talking to customers, putting out fires and doing damage control when something breaks. And depending on how senior you are, there may be managing a budget and managing devs under you. (Then you may get to deal with HR for hiring, firing and performance evaluations.) It's much more rigorous and often very different from the sort of speculative, independant exploratory development that takes place at grad school.
I'm not trying to put down grad school (I wouldn't be back if I didn't think it had value), but someone who's never worked in the commercial sector will lack a lot of the real-life experience that senior engineers there need. And an advanced degree is not a substitute.
Slashdot Mirror
Dr. Gaius Baltar from the new BSG has rapidly become a favorite of mine. He's such a weasel and so much fun (and exhausting) to watch!
Sometime within the last year or two if I recall. Robert Iger took over after Eisner.
Yes, but how do you do the patent search? Typically the patents are phrased in such obscure but broad language that even a professional patent search is unlikely to turn everything up. Plus, paradoxically doing such a search actually increases your liability if you should overlook something. As for designing around things, that might work if so many patents on obvious things weren't granted. Inventors shouldn't have to design around things that are intuitive in their field.
Nah. Nothing so esoteric as that. Just simple courtesy to your fellow man, that's all. That's how most etiquette originated: trying to save each other a bit of trouble. It worked, too, once upon a time.
/misses netiquette.
The reason why 'u' and 'r' started appearing was mostly because of the speed issue.
I've always found the speed argument silly and selfish. Sure it might be slightly faster for the writer to omit letters that are technically redundant from an information theory perspective, but it cost each and every reader more time to "decode" the message. For something that's going to be read by a number of people, the aggregate time savings to everyone is net negative. I've always felt it to be truely selfish and lazy on the part of the writer and shows no consideration for the readers. If the writer doesn't have time to draft a proper message, maybe they shouldn't be posting.
But maybe I'm just a curmudgeon who'd rather see people use the extra time it takes to write properly to think through what they're saying.
x = x++; // add one to x
Not only is it not useful but it's wrong too! The code modifies x twice between sequence points. It'll only add one to x if you're lucky, but it's behaviour is actually undefined in C/C++. "Nasal demons" and all that jazz.
I believe the grand parent was referring to Philips' position of refusing to license their CD-DA logo to discs that don't conform to the Red Book standard. If the logo is absent from the disc, this suggest that it's in non-conformance, likely due to DRM. Even if the logo is present, however, that's no guarantee that it is DRM free, but the chances are better if the logo is there.
Because 3 floats/pixel adds up really quickly. You can do it now, but that was horrendously expensive back in the day. (Heck, 24bpp was considered insanely expensive once upon a time. That's why people used hacks like palettes and indexed images.)
There's also the question of how you flatten that down to the gamut of the monitor. You're typical monitor has maybe 2 to 3 orders of magnitude. Figuring out the best way to compress HDR down to that scale (i.e. tone-mapping) is still an active area of research in computer graphics even though folks here are acting like it's a solved problem.
Lua at least already has something like this. They're called asymmetric coroutines. Also check out Scheme, Scheme's had continuations forever and some folks have done interesting things with them.
I subscribe to Dr. Dobbs Journal (a.k.a DDJ) and the C/C++ Users Journal (CUJ). I liked DDJ a lot when I first started subscribing to it 15 years ago. It got a little lean a few years ago but seems to be getting better lately. Still, though, it's a classic and the CD-Rom archive is terrific. CUJ is more focused on C++, but I like it for keeping up with modern trends in the C++ world.
Uggh... minor orthographic issues I can forgive but this one drives me nuts! Once and for all, the phrase is "for all intents and purposes." Thank you.
I have a quote that I added to my fortune file of a bit in "The Rebel Angels" by Robertson Davies where a character (the head of a huge banking firm) is explaining exactly this to his date:
"Money's something you shove around, like electricity."
"Like electricity?"
"Like large power-grids, and transformers, and that sort of thing. The diffusion of electricity is an extremely important kind of engineering. You decide where to put the energy, and how to get it there, according to the result you expect. Money is a form of power."
-- "The Rebel Angels" Robertson Davies
Yes. I have some soundtracks to games I've never played, but still find the soundtracks quite enjoyable. Of course, these are often from games in a series where I have played some of them, just not all of them. (e.g. Castlevania and Final Fantasy). So I'll admit that I'm predisposed to like them.
Indeed! I always thought that FF4 was the perfect length. Long enough to be satisfying when I beat it, but not so long that I couldn't beat it in a few weeks. FF5 and FF6 were a bit longer but still reasonable. The switch to the PS was about where they lost me.
Part of my problem is that I enjoyed the old FF games for the story. My problem is that if a game starts taking me too long to beat (not having as much free time anymore doesn't help) and tries to cram too much story in, I begin to forget the parts of the game's story before I can finish it. That's not so much fun for someone like me.
I suspect he meant the upcoming Unreal Engine 3. Unreal and UT99 used the Unreal Engine 1. UT2k3 and UT2k4 both used the same basic Unreal Engine 2. UT2k7 will be based on the new Unreal Engine 3. Epic's Unreal Technology Roadmap has the details.
Actually, I find the "other Firefox extensions" link almost more interesting. The Google Suggest extension looks pretty darn slick.
I'd also suggest the red book as a companion/reference. The OpenGL consortium's site also has some some documentation and tutorials (e.g. the spec downloadable in PDF form.)
I'm a "gentile" who moved here to Salt Lake City at the beginning of the year to go back to grad school and this is pretty much how it's looked to me so far. About the only thing I have to add is that I've been told that many people aren't quite as wealthy as they seem at first glance here. The LDS tendency to mary young often leads to people overextending their credit trying to purchase a house too early.
And yeah, I have yet to really meet annoying pushy LDS folks like I have evangelicals. Polite protest of disinterest have worked just fine.
There also turn out to be some surprising advantages to the oddities in legislation here. For example, the recent SCOTUS Eminent Domain ruling has no effect here as the state government recently passed legislation banning eminent domain for redevelopment purposes (except, in a twisted joke, for the redevelopers own property.)
I should mention, however, that my views are somewhat coloured by a fairly centrists view of American politics and religion, so the politics and religion here don't really bother me as much as they might some. (Honestly, for the moment I find them a somewhat refreshing change from my old heavily blue state.)
Yep, that's the way my UT2k4 iCTF clan handles it. We've found most of the anti-cheat mods out there tend to make the game lag horribly and don't even catch many cheaters anyway. So we've taken to just having a large group of admins around to keep an eye on things and hand out bans.
Social problem, social solution...
(Of course, why people still try to cheat when they see players with our tag around is still a mystery to me!)
I've been looking at senior software developer positions, but is that too high up the ladder for someone 'fresh' to cope with?
Depends on where you work. I worked for several years after college and then went back to school full time for an advanced degree. At least in my experience, there's a world of difference between what the senior software engineers did and the kind of development that I do in grad school.
Most academic types don't have to worry about making their code bulletproof, "productizing" it, requirements documents, tech specs, working with UI folks, working with QA folks and bug DBs, or coding to a schedule as part of team. Then there's talking to customers, putting out fires and doing damage control when something breaks. And depending on how senior you are, there may be managing a budget and managing devs under you. (Then you may get to deal with HR for hiring, firing and performance evaluations.) It's much more rigorous and often very different from the sort of speculative, independant exploratory development that takes place at grad school.
I'm not trying to put down grad school (I wouldn't be back if I didn't think it had value), but someone who's never worked in the commercial sector will lack a lot of the real-life experience that senior engineers there need. And an advanced degree is not a substitute.