I think there are relatively new exploits that are neither rootkit nor true virus. They seem to involve corrupting IE and crippling java update and then bring your searches to infected websites to constantly download trojans using a java bug (that is fixed in an update).
Incidentally, I've tried both and Spybot and Malware and both seem to find stuff the other does not (malware seems to be better about tracking cookies, spybot exploits). Unfortunately, neither seems to have found the problem I am currently chasing, which is a combination of crippling java update and exploiting IE to go to malicious websites whenever the search box in IE7 is used.
I've only partially had time to deal with this, but according to netstat, there are no open ports again (after I removed a spambot for the second time). I've removed all viruses from the box (using virus checkers and manually removing one that apparently was just found in the wild a couple days ago). I've checked and the search is set correctly both in IE7 and in the registry. As far as I can tell, the machine is clean again - except when I do a search - it always puts malicious sites first. I'm going to go through IE's registry entries with a fine-tooth comb tonight, and if that fails, reinstall Windows.
In fact, I am working on just such a case. By dormant, I mean the initial infection was removed, but the virus added some changes to IE so searches almost exclusively go to infected websites and exploit a java bug to reinfect the machine.
The PC in question was my wife's, and she had followed a link to an unknown sender's e-card (which happened to arrive on her birthday) and it exploited her gullibility and a java bug to install the trojan XP Antivirus '08. I managed to eradicate that virus, but it made a change to IE that I missed initially that takes searches to infected websites and exploits the java bug again to reinfected the machine (mainly with other viruses - Virtumunde has been the latest - both of these are Russian Federation originating). Antivirus software doesn't catch the infections because they happen in resident memory, but the software does find them after they've written files.
The problem is, she needs to have her java patched to remove the java back door, but the virus seems to have tampered with java and it will not patch. I'm going to try a manual uninstall and reinstall tonight. I also likely need to reinstall IE (will try a registry fix first using my XP box as a reference), but MS has made that impossible by design, so I'll probably need to reinstall the entire OS.
Re:Non-Tech Percent of Web Traffic from Chrome
on
Google Chrome, Day 2
·
· Score: 5, Informative
yeah - I just tested it with a javascript heavy app and it seems very snappy. The same app doesn't even run correctly on Firefox 3 (it does on IE and Firefox 2, and I believe we filed a FF3 bug).
I have not done extensive testing (heck, it won't be supported, so there is no reason to), but it seems a good effort so far.
If you want to know why pensions are dying, look no further than the auto industry - most US auto companies offer(ed) pensions and were not able to compete with Chinese and Japanese companies that do not. It also looks bad from a debt load on the company's bottom line (unless they saved for it, but this is America - since when did anyone save? - they do, however, have to set aside some money). Pensions also put the employer at risk - it counts against their debt and ability to borrow.
401k has no such disadvantage - the entire burden and risk is put on the employee. Employers that still have pensions are those with nearly no risk - the US Government, for instance (can't pay the pensions? raise taxes!). I've had pensions twice, both times in large companies, and by (bad?) luck or coincidence my group was spun off just before the majority of employees vested in those pensions (I was a week away and had to transfer that money to an IRA). Incidentally, both times my company was spun off it was so the horribly unprofitable parent could skim a huge amount of cash off their profitable child so their stock wouldn't go junk (and neither company exists today because they sold off all their profitable groups).
Of course, you could argue that such a feature has existed in most text editors (e.g. vi has cont-D and cont-U) that could be key-bound to a single key, and that they exhibit exactly the same behavior (as long as the terminal window is the size of a printed page, which was default size on the old Solaris boxes I used). I used to keybind all the time with those silly Solaris keyboards with the extra 10 function keys on the side that were never bound for anything but framemaker (names were weird - I think Prior and Again were the ones I used).
MS is arguing that Page Up and Page Down brings you to exactly the same spot in the next page, which is not always typical behavior for that key (as mentioned in Parent, AcroRead goes to the top of the next page). It IS, however, typical behavior for all text editors I've ever used, including some that support the page up and page down keys and existed (long) before 2005 and some that support a "page view" for printing purposes.
Did you read what I said? I said ATI destroys nVidia in SLI (and yes, that is due to architecture issues). Single card vs single card is a different issue entirely, although nVidia tends to gulp a lot more power.
As a minus to ATI, they have horribly inefficient shaders - I have 192 shaders in one card, but my 64 shader nVidia card often outperforms it in shader stress tests (I can tune some of that performance back, but it still isn't getting 192 shaders of performance - maybe 80 nVidia shaders). Power consumption-wise, ATI wins hands down (I think the nVidia card needs a 630 Watt power supply, the ATI a 450).
My personal peeve against ATI, and it's purely as a developer, is driver related - they tend to wait for official ARB on the OpenGL side rather than releasing EXTs like nVidia does - the hardware is usually there already (from the DirectX side) and an EXT means there is a draft design accepted by the majority of the ARB. I write cross-platform, so I need to at least write OpenGL, so it's very frustrating to have to write three sets of code just because ATI cards can't handle the latest features and MS chose not to support new features in DX9. OpenGL and DX9 are fairly similar models, so not too hard to rewrite, but DX10 requires a lot more code work (due to an entirely different model). Sometimes I just throw up my hands and go screw it, this feature isn't supported on ATI cards or DX because it takes too much time to implement. Just to give you an idea, I had to write my last custom shader 5 times (2x for a modern hardware technique in DX10 and OGL and 3x for a fallback to older hardware for DX9, OGL, and a less accurate/less shader intensive DX10 method) - it's just an incredible time sink for something I should at most have had to write twice at most (it runs on the same hardware, dammit).
I'm pretty sure the duration is longer than most matches - the only true Meth addict I've known had some he claimed lasted for 12+ hours and had stayed awake for more than 2 days straight before crashing and sleeping for a day. A high school/college buddy of mine eventually carted that guy off to treatment (in fact, that's how I knew him in the first place - my friend befriended him during the first days of college because they were dorm neighbors and we later played rec soccer, hockey and broomball together). This was before the drug really went mainstream - it was early 1990s.
I've always been under the impression that most hair bands (and grunge bands) did heroin (or speedballs or highballs) not straight coke, though I'm sure some did in some mixture. Coke is, according to an ex-roommate (and rehabbed addict), "the most expensive 15 minute high you can buy," so not much use when you have at least an hour long set. Heroin doesn't affect the motor reflexes and lasts a long time (and destroys the user, and is permanently and physically addictive the first time it's used).
I've unfortunately known musicians that OD'd on drugs in the 1990s (I can't say exactly what killed them because in both cases it was a "cocktail" of mixed drugs including heroin), so I can't say I have much positive to say about drugs or drug culture (especially when "hard" drugs are involved).
As for pot, I've done it once (in my band days), and for me, it made me feel stupid and incompetent (like I was out-of-touch with my body), and I couldn't lose that feeling for more than a week, so personally, aside from hating the experience (I'm WAAAY too control freak for that), I don't see how that could help, but I also know other people that can't function normally without it. I imagine someone that learns to do something while high can only competently do it while high - for instance, I can play guitar right handed, but I'm incompetent if I try to play it left handed - I know what to do, but the motor skills aren't trained right.
ATI and nVidia's latest single card setups are fairly comparable - it's in SLI mode where ATI smokes nVidia.
As for this press release, nVidia has to do this - Intel struck first by sowing a bunch of FUD that directly affects discrete graphics cards, and nVidia's using the exact same argument I used a couple weeks ago here on slashdot - it relies on extremely fast memory and memory bandwidth (in fact, it wouldn't surprise me if that is the bottleneck). In all cases I've seen, they've only shown optimal (highly specular) scenes without photon mapping (one of the cheapest routines for soft shadows in ray tracing). Show me some ray traced soft shadowed human figures in real time that don't look like plastic mannequins (so subsurface scattering) and a highly diffuse scene and I'll be a believer.
Then switch to ATI, where pretty much no OpenGL extension is supported. Need GL_EXT_draw_instanced for _instance_ (ha ha - yeah, ok, that was lame)? You won't find it on an ATI card, even though the same DirectX functionality exists and nVidia's supported it for over 2 years. How about GL_EXT_geometry_shader_4? Not ATI again (not that geo shaders have proven to be all that useful). ATI tends to skip extensions until they reach ARB unless they invented it, which often leaves them a generation or two behind DirectX or nVidia in OpenGL. Any delay in an OpenGL release leaves features in the stone age, which may very well be the reason the ARB was driven to release OGL 3.0 with the rework unfinished - basically get the functions in ARB.
From a developer perspective, where I may be adding features that will not be in production for 2-3 years, adding EXTs in early development is not that big of deal, because they should be in mainstream cards by release time. ATI has great DirectX support, and I applaud them for that, but I find their OpenGL support for new features lacking.
Not quite - *Red* LEDs are made from a Gallium Arsenide-Gallium Phosphide alloy. You're probably thinking of other LEDs like *Green* LEDs, which can be Gallium Nitride or *Blue* LEDs which can be Indium Gallium Nitride. *White* LEDs can be either an RGB combo or a Blue LED with a phosphor.
Gallium Arsenide is used in a lot of the semiconductor industry, and not just in LEDs - it's used in some ICs, solar cells, and laser diodes, as well.
I wouldn't call most short lived - maybe Altair or MSX Basic (remember Microsoft's failed z80 based computer?) - MS released BASIC for every successful platform - Apple (Applesoft BASIC), Tandy (Color BASIC, TRS-80 Level II BASIC), IBM (too many to count - QBASIC and Visual I remember - QBASIC was on OS/2 as well), Commodore (Commodore BASIC), Atari (heh - I read TFA to learn that one), and Amiga (um... damn, can't remember the name, but I'm certain it was MS or MS derived).
Personally, I only used the Atari 400/800 a couple of times. I don't remember them even offering a floppy drive for that one (I remember one later model having them built-in). Shows what I know. I do remember magazines with semi-cross platform BASIC code, which was fun in my pre-teens.
One of the killers for the 400/800 was also the expensive and heavy Faraday Cage they were forced to include by FCC ruling to avoid RF leakage. Apple got around it by licensing an inexpensive FCC approved RF modulator from a separate company (see here) to be sold separately with new Apple ][s (and Commodore avoided the issue, I believe - the PET was grandfathered and the Vic 20 had time to learn by Apple's example). In a low end price war between the 400 and Vic 20 (and later C64), the 400 just couldn't compete.
Heh - neither, because the furniture is actually pronounced just like the meal, buff ay.
The ett part of the name is either pronounced 'it' or 'ette' (like the French word ending) usually.
In response to the grandparent, those are estimated numbers, so in reality any one of them is the true richest man at any particular moment (yes, man - the world's richest woman is reported to be L'Oreal heir Liliane Bettencourt at a paltry 23 billion, but that may entirely depend on pre-nuptials).
Also part of Anarchy: Online and the in development The Secret World.
Incidentally, Sid Meier did mostly sims (the two I owned were F15 Strike Eagle and Silent Service) and his only real hit before Civilization was Sid Meier's Pirates! (which bore his name in a marketing move). I don't think anyone could have named Sid Meier before they put his name on the Pirates! box, but after Civ, nobody forgot him.
The Longest Journey is probably the best adventure game I've ever played. Dreamfall was not (decent plot, but gameplay and length issues). Most adventure games have strolled into the 'Myst' model, which is they're mostly about the puzzles. TLJ was all about the characters, which, to be honest, is the only type of Adventure game I like. For instance, I disliked Syberia, another female driven adventure game - the character was flat and the plot boring (it had decent puzzles - oh, yay). I play a lot of RPGs, as well, and RPGs have, in fact gone towards a character driven model, but even games like Mass Effect haven't gotten me to really care about the characters (spoiler alert) - Ashley or Kaidan (possibly) dying? No big deal. There is nothing more for them - their cycle is complete. As a plot character, Ashley was a bit more developed, but still seemed incomplete - she chats with her sister but if she dies, do you have to break the bad news (or delegate it)? No. How about that as a romance possibility - consoling Ashley's sister after breaking the bad news? It certainly would be a variation - Bioware has used the same romance model since Baldur's Gate 2 with no variation and to me it's dull and over-trodden. They should go back and play Planescape: Torment (which is the first game to use that model, I believe - the jealousy between the Tiefling and the Succubus was classic, though to be quite honest, the game was a bit too talky for me in a "click through 8 pages of meaningless dialog" sort of way).
Not to mention it was NOT a shooter, which Wolfenstein 3D and onward were - it is considered the first action/stealth game. Beyond Castle Wolfenstein even let you drag corpses around.
Castle Wolfenstein was also one of the first games to have a major mod, Castle Smurfenstein, by Dead Smurf Software. I enjoyed that one better than Dino Smurfs (a mod on Dino Eggs, which was not my favorite game to begin with - I loved Wolf and hate those little blue bastards).
Silas Warner would be rolling in his grave if he knew of the ignorance on this thread;)
hmm - I was a bit underwhelmed, myself. First of all, graphics people have done GPU ray tracing on small static scenes has been around since about 2003 or 2004. The real limitation then was available memory. This is a standard ray tracer (no photon mapping), and highlighting an optimal model (a reflective one, not a diffuse one). They mention it scales linearly, which means they replaced the fixed function pipeline and used only shaders.
The scene shown is 2 million polygons (about the baseline of this generation's hardware) but with a single light source and hard shadows. Three reflections is adequate to get a decent looking scene (I've done good looking scenes with 2 reflections, but it depends on the scene).
Even the gameplay movie they have up on their site (from E3) only shows monsters dropping orbs when they die. The way it currently looks, bosses have certain attacks and you need to run around a lot and avoid them and pick up orbs dropped by previously beaten monsters.
Back to listening to "Town" and "Tristram" by Matt Uelmen on B-net mp3 player - brings back memories (always did like the Bauhaus [specifically the end of Mask] sound of those tracks with the 12 string - all the new stuff released so far for Diablo 3 sounds more Dead Can Dance, but some of the earlier tracks also sounded like DCD).
you mean they got BORED in a meeting (don't we all?) and decided to do something productive during that big 5 hour power point presentation. That or slow news day. Or maybe the nukes are on the way and this is their little way of trying to humor us before the end - I don't keep up with that kind of stuff.
A friend I had over at Loki said that lots of people said they wanted Linux versions, but when they actually did the ports, the market wasn't there.
The key seems to also be to release concurrently if possible. With Linux or even MacOS you either need to do release concurrently or get the port out before it works with Wine or Cedega or people just buy the Windows version (mac gamers will dual boot or run Windows or Linux in Parallels - I have an acquaintance that runs Parallels-Linux-WINE on a macbook pro because he refuses to run Windows and loves Counterstrike).
No, the problem is more that most of the players racked up massive debt during the.com boom and due to several factors including the erosion of customers in their land phone base and loss of long distance revenue, they have lacked money to dump into infrastructure. Every single one of the has a massive debt.
Verizon with one of the smallest networks in the most populous part of the country invested in more infrastructure, despite its debt (something like 40 billion). AT&T merged and merged and the other player, Qwest, bought out US West (jokingly called US Worst by some customers, like me - terrible service, terrible DSL speeds priced 2-5x competition, and terrible tech support) but then had to ditch their lucrative Yellow Pages to avoid going junk in 2002.
With 50Mbit service at around $150 (+ fees and tax), 200Mbps service would be over $600, so no, we'd have to pay about 30x that or about $7200 per year compared to $240 (plus WoW fees...).
I'd have to agree - DX10 was mostly an under the hood update and redesign (and 10-20% speed boost, offset by Vista's slowness - Vista itself is written on top of a special version of DX9, not 10), and while it does include a few cool features, writing code to include Geometry Shaders for instance, requires a complete redesign of the shader in most cases, and since that shader is only usable on Vista (as opposed to the OpenGL one which is usable everywhere the hardware exists), it's just extra effort for little substantial gain.
Personally, I think Geometry Shaders are useful, but mainly for shaving renderpasses when doing physics simulations in hardware (you don't need to cram the data into a texture - yay!). I'm not sure why Carmack believes they were only/mainly for stencil shadow volume, although he did have a lot to do with improving that technique and probably has a lot more contact with IHVs than I do - I thought they were a first crack at creating a tessellation shader (but they don't do that well due to vertex emission limitations).
Of course they render in CPU - Ray Tracing (and similar) need access to the entire scene to properly render reflections, color bleed, etc. and storing this in GPU RAM is laughable (even the 32MB FireGL card couldn't handle). Most CAD and CAD-like rendering packages do modeling with polygons for positioning and drafting, then switch to Ray Tracing for the actual render, which is done in system memory.
And blaming the slip and lack of features on CAD companies as per the original post is silly, too - many still write to a baseline 1.1 spec (some may add features, but it's gotta still work with 1.1). The promise for 3.0 was to freeze the fixed function pipeline and create profiles like the "Lean and Mean" profile to work with newer systems - I don't see why anyone would have a problem with that - basically you use the old way by default or you set a profile with a call like GL_USE_PROFILE(LEAN_AND_MEAN); (function name is made up) when setting up to get the new way and later release a streamlined API without backward compatibility (but intended to run side-by-side with OpenGL 3). My understanding is that was what they were doing when they "hit some problems" and then went silent last year. These problems are likely graphical hardware companies crying foul about the intention to have 2 APIs they have to support.
The reality is, 3.0 was always intended as mostly an "under the hood" release, where they deprecated old APIs and did a bunch of prep work for future streamlining that was not possible while the old APIs still existed. The disappointment is the slip in certain features like the object model (which is still coming according to some people on #opengl on IRC).
What really bugged me about this whole debacle is Khronos was very open with what they were doing and when they were doing it and then suddenly stopped talking. Even the newsletter had no real progress details - they just stopped talking and kept programmers hopes up. They could have handled this in a much better way.
Slashdot Mirror
I think there are relatively new exploits that are neither rootkit nor true virus. They seem to involve corrupting IE and crippling java update and then bring your searches to infected websites to constantly download trojans using a java bug (that is fixed in an update).
Incidentally, I've tried both and Spybot and Malware and both seem to find stuff the other does not (malware seems to be better about tracking cookies, spybot exploits). Unfortunately, neither seems to have found the problem I am currently chasing, which is a combination of crippling java update and exploiting IE to go to malicious websites whenever the search box in IE7 is used.
I've only partially had time to deal with this, but according to netstat, there are no open ports again (after I removed a spambot for the second time). I've removed all viruses from the box (using virus checkers and manually removing one that apparently was just found in the wild a couple days ago). I've checked and the search is set correctly both in IE7 and in the registry. As far as I can tell, the machine is clean again - except when I do a search - it always puts malicious sites first. I'm going to go through IE's registry entries with a fine-tooth comb tonight, and if that fails, reinstall Windows.
In fact, I am working on just such a case. By dormant, I mean the initial infection was removed, but the virus added some changes to IE so searches almost exclusively go to infected websites and exploit a java bug to reinfect the machine.
The PC in question was my wife's, and she had followed a link to an unknown sender's e-card (which happened to arrive on her birthday) and it exploited her gullibility and a java bug to install the trojan XP Antivirus '08. I managed to eradicate that virus, but it made a change to IE that I missed initially that takes searches to infected websites and exploits the java bug again to reinfected the machine (mainly with other viruses - Virtumunde has been the latest - both of these are Russian Federation originating). Antivirus software doesn't catch the infections because they happen in resident memory, but the software does find them after they've written files.
The problem is, she needs to have her java patched to remove the java back door, but the virus seems to have tampered with java and it will not patch. I'm going to try a manual uninstall and reinstall tonight. I also likely need to reinstall IE (will try a registry fix first using my XP box as a reference), but MS has made that impossible by design, so I'll probably need to reinstall the entire OS.
yeah - I just tested it with a javascript heavy app and it seems very snappy. The same app doesn't even run correctly on Firefox 3 (it does on IE and Firefox 2, and I believe we filed a FF3 bug).
I have not done extensive testing (heck, it won't be supported, so there is no reason to), but it seems a good effort so far.
If you want to know why pensions are dying, look no further than the auto industry - most US auto companies offer(ed) pensions and were not able to compete with Chinese and Japanese companies that do not. It also looks bad from a debt load on the company's bottom line (unless they saved for it, but this is America - since when did anyone save? - they do, however, have to set aside some money). Pensions also put the employer at risk - it counts against their debt and ability to borrow.
401k has no such disadvantage - the entire burden and risk is put on the employee. Employers that still have pensions are those with nearly no risk - the US Government, for instance (can't pay the pensions? raise taxes!). I've had pensions twice, both times in large companies, and by (bad?) luck or coincidence my group was spun off just before the majority of employees vested in those pensions (I was a week away and had to transfer that money to an IRA). Incidentally, both times my company was spun off it was so the horribly unprofitable parent could skim a huge amount of cash off their profitable child so their stock wouldn't go junk (and neither company exists today because they sold off all their profitable groups).
Of course, you could argue that such a feature has existed in most text editors (e.g. vi has cont-D and cont-U) that could be key-bound to a single key, and that they exhibit exactly the same behavior (as long as the terminal window is the size of a printed page, which was default size on the old Solaris boxes I used). I used to keybind all the time with those silly Solaris keyboards with the extra 10 function keys on the side that were never bound for anything but framemaker (names were weird - I think Prior and Again were the ones I used).
MS is arguing that Page Up and Page Down brings you to exactly the same spot in the next page, which is not always typical behavior for that key (as mentioned in Parent, AcroRead goes to the top of the next page). It IS, however, typical behavior for all text editors I've ever used, including some that support the page up and page down keys and existed (long) before 2005 and some that support a "page view" for printing purposes.
Did you read what I said? I said ATI destroys nVidia in SLI (and yes, that is due to architecture issues). Single card vs single card is a different issue entirely, although nVidia tends to gulp a lot more power.
As a minus to ATI, they have horribly inefficient shaders - I have 192 shaders in one card, but my 64 shader nVidia card often outperforms it in shader stress tests (I can tune some of that performance back, but it still isn't getting 192 shaders of performance - maybe 80 nVidia shaders). Power consumption-wise, ATI wins hands down (I think the nVidia card needs a 630 Watt power supply, the ATI a 450).
My personal peeve against ATI, and it's purely as a developer, is driver related - they tend to wait for official ARB on the OpenGL side rather than releasing EXTs like nVidia does - the hardware is usually there already (from the DirectX side) and an EXT means there is a draft design accepted by the majority of the ARB. I write cross-platform, so I need to at least write OpenGL, so it's very frustrating to have to write three sets of code just because ATI cards can't handle the latest features and MS chose not to support new features in DX9. OpenGL and DX9 are fairly similar models, so not too hard to rewrite, but DX10 requires a lot more code work (due to an entirely different model). Sometimes I just throw up my hands and go screw it, this feature isn't supported on ATI cards or DX because it takes too much time to implement. Just to give you an idea, I had to write my last custom shader 5 times (2x for a modern hardware technique in DX10 and OGL and 3x for a fallback to older hardware for DX9, OGL, and a less accurate/less shader intensive DX10 method) - it's just an incredible time sink for something I should at most have had to write twice at most (it runs on the same hardware, dammit).
I'm pretty sure the duration is longer than most matches - the only true Meth addict I've known had some he claimed lasted for 12+ hours and had stayed awake for more than 2 days straight before crashing and sleeping for a day. A high school/college buddy of mine eventually carted that guy off to treatment (in fact, that's how I knew him in the first place - my friend befriended him during the first days of college because they were dorm neighbors and we later played rec soccer, hockey and broomball together). This was before the drug really went mainstream - it was early 1990s.
I've always been under the impression that most hair bands (and grunge bands) did heroin (or speedballs or highballs) not straight coke, though I'm sure some did in some mixture. Coke is, according to an ex-roommate (and rehabbed addict), "the most expensive 15 minute high you can buy," so not much use when you have at least an hour long set. Heroin doesn't affect the motor reflexes and lasts a long time (and destroys the user, and is permanently and physically addictive the first time it's used).
I've unfortunately known musicians that OD'd on drugs in the 1990s (I can't say exactly what killed them because in both cases it was a "cocktail" of mixed drugs including heroin), so I can't say I have much positive to say about drugs or drug culture (especially when "hard" drugs are involved).
As for pot, I've done it once (in my band days), and for me, it made me feel stupid and incompetent (like I was out-of-touch with my body), and I couldn't lose that feeling for more than a week, so personally, aside from hating the experience (I'm WAAAY too control freak for that), I don't see how that could help, but I also know other people that can't function normally without it. I imagine someone that learns to do something while high can only competently do it while high - for instance, I can play guitar right handed, but I'm incompetent if I try to play it left handed - I know what to do, but the motor skills aren't trained right.
ATI and nVidia's latest single card setups are fairly comparable - it's in SLI mode where ATI smokes nVidia.
As for this press release, nVidia has to do this - Intel struck first by sowing a bunch of FUD that directly affects discrete graphics cards, and nVidia's using the exact same argument I used a couple weeks ago here on slashdot - it relies on extremely fast memory and memory bandwidth (in fact, it wouldn't surprise me if that is the bottleneck). In all cases I've seen, they've only shown optimal (highly specular) scenes without photon mapping (one of the cheapest routines for soft shadows in ray tracing). Show me some ray traced soft shadowed human figures in real time that don't look like plastic mannequins (so subsurface scattering) and a highly diffuse scene and I'll be a believer.
Then switch to ATI, where pretty much no OpenGL extension is supported. Need GL_EXT_draw_instanced for _instance_ (ha ha - yeah, ok, that was lame)? You won't find it on an ATI card, even though the same DirectX functionality exists and nVidia's supported it for over 2 years. How about GL_EXT_geometry_shader_4? Not ATI again (not that geo shaders have proven to be all that useful). ATI tends to skip extensions until they reach ARB unless they invented it, which often leaves them a generation or two behind DirectX or nVidia in OpenGL. Any delay in an OpenGL release leaves features in the stone age, which may very well be the reason the ARB was driven to release OGL 3.0 with the rework unfinished - basically get the functions in ARB.
From a developer perspective, where I may be adding features that will not be in production for 2-3 years, adding EXTs in early development is not that big of deal, because they should be in mainstream cards by release time. ATI has great DirectX support, and I applaud them for that, but I find their OpenGL support for new features lacking.
Not quite - *Red* LEDs are made from a Gallium Arsenide-Gallium Phosphide alloy. You're probably thinking of other LEDs like *Green* LEDs, which can be Gallium Nitride or *Blue* LEDs which can be Indium Gallium Nitride. *White* LEDs can be either an RGB combo or a Blue LED with a phosphor.
Gallium Arsenide is used in a lot of the semiconductor industry, and not just in LEDs - it's used in some ICs, solar cells, and laser diodes, as well.
I wouldn't call most short lived - maybe Altair or MSX Basic (remember Microsoft's failed z80 based computer?) - MS released BASIC for every successful platform - Apple (Applesoft BASIC), Tandy (Color BASIC, TRS-80 Level II BASIC), IBM (too many to count - QBASIC and Visual I remember - QBASIC was on OS/2 as well), Commodore (Commodore BASIC), Atari (heh - I read TFA to learn that one), and Amiga (um... damn, can't remember the name, but I'm certain it was MS or MS derived).
Personally, I only used the Atari 400/800 a couple of times. I don't remember them even offering a floppy drive for that one (I remember one later model having them built-in). Shows what I know. I do remember magazines with semi-cross platform BASIC code, which was fun in my pre-teens.
One of the killers for the 400/800 was also the expensive and heavy Faraday Cage they were forced to include by FCC ruling to avoid RF leakage. Apple got around it by licensing an inexpensive FCC approved RF modulator from a separate company (see here) to be sold separately with new Apple ][s (and Commodore avoided the issue, I believe - the PET was grandfathered and the Vic 20 had time to learn by Apple's example). In a low end price war between the 400 and Vic 20 (and later C64), the 400 just couldn't compete.
Heh - neither, because the furniture is actually pronounced just like the meal, buff ay.
The ett part of the name is either pronounced 'it' or 'ette' (like the French word ending) usually.
In response to the grandparent, those are estimated numbers, so in reality any one of them is the true richest man at any particular moment (yes, man - the world's richest woman is reported to be L'Oreal heir Liliane Bettencourt at a paltry 23 billion, but that may entirely depend on pre-nuptials).
The question is, why go that far?
yeah - why bother when Microsoft will just deprecate it and replace it with proprietary underbar functions anyway?
Dreamfall - did you read TFA?
Also part of Anarchy: Online and the in development The Secret World.
Incidentally, Sid Meier did mostly sims (the two I owned were F15 Strike Eagle and Silent Service) and his only real hit before Civilization was Sid Meier's Pirates! (which bore his name in a marketing move). I don't think anyone could have named Sid Meier before they put his name on the Pirates! box, but after Civ, nobody forgot him.
The Longest Journey is probably the best adventure game I've ever played. Dreamfall was not (decent plot, but gameplay and length issues). Most adventure games have strolled into the 'Myst' model, which is they're mostly about the puzzles. TLJ was all about the characters, which, to be honest, is the only type of Adventure game I like. For instance, I disliked Syberia, another female driven adventure game - the character was flat and the plot boring (it had decent puzzles - oh, yay). I play a lot of RPGs, as well, and RPGs have, in fact gone towards a character driven model, but even games like Mass Effect haven't gotten me to really care about the characters (spoiler alert) - Ashley or Kaidan (possibly) dying? No big deal. There is nothing more for them - their cycle is complete. As a plot character, Ashley was a bit more developed, but still seemed incomplete - she chats with her sister but if she dies, do you have to break the bad news (or delegate it)? No. How about that as a romance possibility - consoling Ashley's sister after breaking the bad news? It certainly would be a variation - Bioware has used the same romance model since Baldur's Gate 2 with no variation and to me it's dull and over-trodden. They should go back and play Planescape: Torment (which is the first game to use that model, I believe - the jealousy between the Tiefling and the Succubus was classic, though to be quite honest, the game was a bit too talky for me in a "click through 8 pages of meaningless dialog" sort of way).
Not to mention it was NOT a shooter, which Wolfenstein 3D and onward were - it is considered the first action/stealth game. Beyond Castle Wolfenstein even let you drag corpses around.
Castle Wolfenstein was also one of the first games to have a major mod, Castle Smurfenstein, by Dead Smurf Software. I enjoyed that one better than Dino Smurfs (a mod on Dino Eggs, which was not my favorite game to begin with - I loved Wolf and hate those little blue bastards).
Silas Warner would be rolling in his grave if he knew of the ignorance on this thread ;)
hmm - I was a bit underwhelmed, myself. First of all, graphics people have done GPU ray tracing on small static scenes has been around since about 2003 or 2004. The real limitation then was available memory. This is a standard ray tracer (no photon mapping), and highlighting an optimal model (a reflective one, not a diffuse one). They mention it scales linearly, which means they replaced the fixed function pipeline and used only shaders.
The scene shown is 2 million polygons (about the baseline of this generation's hardware) but with a single light source and hard shadows. Three reflections is adequate to get a decent looking scene (I've done good looking scenes with 2 reflections, but it depends on the scene).
Even the gameplay movie they have up on their site (from E3) only shows monsters dropping orbs when they die. The way it currently looks, bosses have certain attacks and you need to run around a lot and avoid them and pick up orbs dropped by previously beaten monsters.
Back to listening to "Town" and "Tristram" by Matt Uelmen on B-net mp3 player - brings back memories (always did like the Bauhaus [specifically the end of Mask] sound of those tracks with the 12 string - all the new stuff released so far for Diablo 3 sounds more Dead Can Dance, but some of the earlier tracks also sounded like DCD).
you mean they got BORED in a meeting (don't we all?) and decided to do something productive during that big 5 hour power point presentation. That or slow news day. Or maybe the nukes are on the way and this is their little way of trying to humor us before the end - I don't keep up with that kind of stuff.
A friend I had over at Loki said that lots of people said they wanted Linux versions, but when they actually did the ports, the market wasn't there.
The key seems to also be to release concurrently if possible. With Linux or even MacOS you either need to do release concurrently or get the port out before it works with Wine or Cedega or people just buy the Windows version (mac gamers will dual boot or run Windows or Linux in Parallels - I have an acquaintance that runs Parallels-Linux-WINE on a macbook pro because he refuses to run Windows and loves Counterstrike).
actually, it's also why New York (and parts of New Jersey) and Virginia pay $90 for 50Mbps and the rest of us pay $140-150 + fees.
In some parts of the country you pay $100+ for 6-10Mbit service.
No, the problem is more that most of the players racked up massive debt during the .com boom and due to several factors including the erosion of customers in their land phone base and loss of long distance revenue, they have lacked money to dump into infrastructure. Every single one of the has a massive debt.
Verizon with one of the smallest networks in the most populous part of the country invested in more infrastructure, despite its debt (something like 40 billion). AT&T merged and merged and the other player, Qwest, bought out US West (jokingly called US Worst by some customers, like me - terrible service, terrible DSL speeds priced 2-5x competition, and terrible tech support) but then had to ditch their lucrative Yellow Pages to avoid going junk in 2002.
Basically, we now have this picture in the US: http://arstechnica.com/news.ars/post/20080424-qwest-finally-gets-in-the-fiber-game-with.html
With 50Mbit service at around $150 (+ fees and tax), 200Mbps service would be over $600, so no, we'd have to pay about 30x that or about $7200 per year compared to $240 (plus WoW fees...).
GLIDE was Open Sourced just before 3dfx was bought by nVidia. http://glide.sourceforge.net/
Feel free to port it to other cards ;)
I'd have to agree - DX10 was mostly an under the hood update and redesign (and 10-20% speed boost, offset by Vista's slowness - Vista itself is written on top of a special version of DX9, not 10), and while it does include a few cool features, writing code to include Geometry Shaders for instance, requires a complete redesign of the shader in most cases, and since that shader is only usable on Vista (as opposed to the OpenGL one which is usable everywhere the hardware exists), it's just extra effort for little substantial gain.
Personally, I think Geometry Shaders are useful, but mainly for shaving renderpasses when doing physics simulations in hardware (you don't need to cram the data into a texture - yay!). I'm not sure why Carmack believes they were only/mainly for stencil shadow volume, although he did have a lot to do with improving that technique and probably has a lot more contact with IHVs than I do - I thought they were a first crack at creating a tessellation shader (but they don't do that well due to vertex emission limitations).
Of course they render in CPU - Ray Tracing (and similar) need access to the entire scene to properly render reflections, color bleed, etc. and storing this in GPU RAM is laughable (even the 32MB FireGL card couldn't handle). Most CAD and CAD-like rendering packages do modeling with polygons for positioning and drafting, then switch to Ray Tracing for the actual render, which is done in system memory.
And blaming the slip and lack of features on CAD companies as per the original post is silly, too - many still write to a baseline 1.1 spec (some may add features, but it's gotta still work with 1.1). The promise for 3.0 was to freeze the fixed function pipeline and create profiles like the "Lean and Mean" profile to work with newer systems - I don't see why anyone would have a problem with that - basically you use the old way by default or you set a profile with a call like GL_USE_PROFILE(LEAN_AND_MEAN); (function name is made up) when setting up to get the new way and later release a streamlined API without backward compatibility (but intended to run side-by-side with OpenGL 3). My understanding is that was what they were doing when they "hit some problems" and then went silent last year. These problems are likely graphical hardware companies crying foul about the intention to have 2 APIs they have to support.
The reality is, 3.0 was always intended as mostly an "under the hood" release, where they deprecated old APIs and did a bunch of prep work for future streamlining that was not possible while the old APIs still existed. The disappointment is the slip in certain features like the object model (which is still coming according to some people on #opengl on IRC).
What really bugged me about this whole debacle is Khronos was very open with what they were doing and when they were doing it and then suddenly stopped talking. Even the newsletter had no real progress details - they just stopped talking and kept programmers hopes up. They could have handled this in a much better way.