I don't think code reuse is an issue for TA. The original developers have split up. Furthermore, Cavedog, the company that made TA, was an experimental effort by some Humongous Entertainment (HE) employees, and operated under HE. HE primarily makes (some very good) children's games.
Who knows what bugs already existed when it was released (who knows that will tell us, anyway)?
Since it's not actively maintained, who knows what other bugs it might have (or expose in the current version of DirectX)?
Since nobody from Cavedog or HE would help look at the TA code, it seems unlikely that another company would want to mess with it at all.
If you were writing an modern adventure game, would you try to reuse code from Zork?
The FFT doesn't count. It's math, not software. It's "just" a clever matrix factorization, not a huge software system. If you include the fft, you might as well go back to 3000BC and the decimal number system, or closer to 500AD and the discovery of how useful "zero" is.
Lapack is much closer to being impressive, but again it's "just" a math library. I'm a math grad student, and I appreciate the FFT and Lapack. But as software goes its longevity isn't as surprising as that of SABRE.
-Paul Komarek
I grew up in Spokane, WA, outside of Fairchild AFB. At the time, they had B-52 bombers there. I remember seeing a picture in the newspaper of about 8 people (2 per side, or something like that) in uniform carrying a large flat chunk of metal across a field. It was part of a flap that had broken off a B-52.
That "part of a flap" had about the same area (projected to 2d) as a Cadillac. Seems like the newspaper said that there were known flap problems at the time, and having flap parts fall off was a somewhat regular event. Thankfully Fairchild AFB is not particularly close to any population centers.
-Paul Komarek
Okay, I'm about to go really off the main topic. But about buying first generation equipment:
There is a small collection of old computer equipment behind some glass in a hallway of CMU's new computer science building. There is a picture of some long-forgotten computer being unloaded from a large truck, back in the early, early days (i.e. just after the Univac changed everything in the US computing industry). They give some advice to the reader, which they gleaned from : "Never buy serial number 1 of *anything*".
The real problem here is figuring out what people mean by a "PC". I agree that the Xbox is not, in the strict sense, a PC. Then again, what is a PC these days? What is the most recent PC standard? The IBM PC standard wasn't even sacred for very long (i.e. once the BIOS was reverse engineered).
However, in the sense that "PC" means computer running windows (I really hate it when people do that), then I suppose the Xbox is mostly like a PC. If PC means being afflicted by legacy, then the Xbox's use of the x86 ISA places it firmly in the aflicted camp. If PC means endorsed by Microsoft, or compliant with whatever Microsoft comes up with today ("Designed for Microsoft Windows Whatever"), then it's a PC.
I guess the real issue on people's minds is how much work is necessary to port from the Xbox what we normally call an x86-based PC. I'm guessing that it would take a lot of work.
I don't know if it's any more complex electrically, but certainly it is mechanically. The Xbox motherboard reminds me of the circuitboards I found as a kid when disassembling broken appliances from the scrapyard. Looking at it makes me want to get the soldering gun out and pull those valuable capacitors to put in my own projects!
I agree with your assesment. The Gamecube strategy makes more sense when you look at it as a very powerful piece of "embedded" hardware designed to run (soft) real-time applications. It provides predictable, mostly on-time behavior, where the predictability comes from the simplicity. Thinking of the Gamecube as a weak general purpose computer is probably a mistake, or at least useless.
I think there's another point to be made here: Nintendo understands that fun games and "high-performance" are orthogonal. Think of how many fun games require a piece of cardboard with an array of alternatingly-colored squares. That cardboard has *really bad computational performance*, but that doesn't hurt it any. And besides, that cardboard is probably the most stable and lowest-power gaming platform around.
You've just inspired another thought for how to use a GBA linked to a Gamecube. How about something like a murder mystery. Without private displays, this type of game would be practically impossible.
Nintendo has always been somewhat enigmatic. Consider that the Gamecube is the first Nintendo console to abandon cartridges. I remember people suggesting that the N64 would die because of the limits cartridges imposed. Maybe Nintendo just understands the demographics better than Sony or Microsoft (or most of us) do, and in particular understands what makes things fun.
Mario Cart with four people in the same room has got to be way more fun than Mario Cart would be online. I know that Quake is way more fun when the room temp has soared to 85F (despite the window and box fans you've set up), you're bumping elbows with your opponents, and people answer when you shout "did you see how far my head flew?" (followed by someone using the toaster or microwave, blowing a fuse deep in the basement of your ancient house, thereby knocking all the computers and lights out).
Re:Proprietary Exchange-support to follow
on
Evolution 1.0 Released
·
· Score: 3, Insightful
The Mozilla guys had a good explanation of why 1.0 is special, at least for them. 1.0 means API freeze -- in the sense that the 1.0 API will be available at least until 2.0 (though it could be extended). This is not so important for users, but for developers outside the Mozilla project, it is critical.
Since Evolution has a plugin facility, maybe the 1.0 obsession exists for similar reasons. Now developers can target it with peace of mind.
Ticketmaster -- agreed! My knowledge of Microsoft networking comes mainly down to my experience, and various whitepapers at Carnegie Mellon University about why they disallow several Microsoft network services. Active Directory is a new thing for MS, and you're right that I'm ignorant on that account. But I have good reason to "think so little of Microsoft's accumen". They've stunk up networking for years.
As for being one of the most targetted networks, I'd agree. However, I'm not so sure they're number 2. I expect that other governments, Yahoo, and places like CERT (hosted by Carnegie Mellon) are also big targets.
As for Adam Smith, he *was* wrong. That's why the US (and every other 'capitalist' economy) uses a regulated implementation of capitalism. The free market makes its decisions based on marketing. If it made decisions based on research and development, then R&D would get 50% of revenues and marketing would get the 5% that R&D gets now.
Where is Microsoft today? Using hundreds of programmers to slowly reinvent unix networking. Why slowly? Because they're waiting for people to forget all the FUD they've put out about how bad unix networking is.
And I don't care about Microsoft's 2 million object directory. Why? Because I can piss farther than you!
While most of what you say sounds reasonable, one thing really caught my eye: "only a few companies have people with enough experience with huge, varied networks". The problem with Microsoft is that they only have experience with huge, homogenous networks; they were blindsided by the internet; they thought remote admin was a bad idea until recently; their network hacks (netbios, for instance) stink on large networks.
I think Microsoft is very *unlikely* to have much useful exerperience with "huge, varied networks". What really gets me is that they seem to *like it this way*.
On the other hand, if they manage to get the story down to 15 minutes, the remaining 75 minutes are really going to suck. And whose going to pay $8 for a 15 minute movie?
I think people have forgotten just how bad the Metallica (was his name Lars or something?) phone interview went. The Metallica guy came off as a complete dick, ass, and idiot. I believe that interview was transcribed more-or-less exactly, and several people commented that Lars or whoever really did sound that dumb in "real life".
When people sit down and write, they're much more careful than when they talk. Oral and literate styles are very different. When someone answers the questions via email, they have a chance to reread the questions and edit their responses, possibly doing so multiple times. Yet some people still screw it up, like the Bush campaign in their slashdot interview.
If we've forgotten just how bad it can be, then we are definitely spoiled brats. I'm glad Bruce cared enough to make the phone call, and that Chris did the best he could at the time.
While I agree that waiting a month wouldn't be so bad, I'd like to say thanks to Chris for giving it his best shot. So maybe he made a bad decision, but at least he tried to make the best of an apparently awkward situation.
I think the reason we wouldn't mind waiting a(nother) month is that we've gotten used to it. I simply forget about the interview once the questions are posted, and am later delighted to find the responses!
I don't think they "disapprove of our way of life in the strongest way." I think they disapprove of our foreign policy. It's not to hard to extend that to disapproving of our collective apathy with respect to our government's foreign policy. All in all, I doubt they care much about us eating twinkies and watching sitcoms.
Marx predicted the dominance of capitalism. I think it's Lenin you have a problem with. I believe Marx even wrote a letter to Lenin explaining why Lenin's perversion of Marx's writings were a bad idea. In fact, Marx's writings are probably more relevant than ever as we watch the battle against corporatism unfold (think "Battle of Seattle" and such).
Whatever one might think of Richard Stallman, it is clear that he causes people to sit up and think. Every time there is an article about RMS on slashdot, an long debate follows. That's quite a contribution for one guy who probably doesn't read slashdot. He makes people think; obviously not everyone who posts is thinking, but at least a few are.
Anyone else notice that L-wave's submission takes a sentence directly from The Register's John Lettice? Look at http://www.theregister.co.uk/content/4/22928.html.
Even if you do use original packaging including molded styrofoam, those clever UPS folks will find a way to damage your shipment. While the damage is likely to be less than what this guy photographed, it's still damage.
We received a rackmount APC Uninteruptable Power Supply (overloaded abbreviations!) via UPS ground. The front plate is steel, about 5mm (about 3/16") thick. It was severaly bent, and we had to straighten it with a benchtop vice.
I've shipped a computer in the original packaging via UPS ground. When it arrived, some case screws had popped out, the case frame was bent, and the plastic front had broken pieces. Needless to say the case didn't really fit right after that.
I've never had damage with FedEx ground, though I've had two computers shipped that way (in original packaging). I had to pick one up from the FedEx distribution location because they messed up the delivery (never put it on the truck), and the box was banged up a little, but the case was fine.
Whenever a vendor doesn't offer FedEx ground, I write a note requesting they consider it.
You're fine. Never use an apostrophe with a possessive pronoun like "its". The original guy is whacked. I've never been completely comfortable with "it's" for "it has" in written language. I'm really uncomfortable with "who's" in written language. Obviously both are spoken.
I'll put in a plug for Python. It's clean, cross-platform, easy to learn (a core focus of the language is making things easy for nonprogrammers), and has support for QT, GTK, Tk, and some other widget libs. There are good books available for learning Python, and the Python community is fiarly strong.
Python is interpreted, so you'll won't see a lot of speed from it. Most applications don't actually need much speed, especially gui apps. However, you wouldn't want to built the SETI-at-home back end in Python, for example.
Microway's prices on the website are horribly out of date. At the beginning of this year, we bought two dual 264DP machines with 4GB each in rackmount cases with slide for $13,500 each. I'm sure the "loaded" config is less now.
That's still expensive, but if you need a *single* *fast* cpu, a bunch of dual P3-1GHz won't do it for you. We need the large virtual memory, and more than that, our code is single-process and single-threaded. We just aren't into clustering, primarily for historical reasons (large, old codebase among others). Other labs might do things differently -- deal with their own memory allocators to span processes, and handle the extreme NUMA-ishness of a cluster. We'd rather put our money up front and save time.
We've got a bigger machine which is basically a 264DP with 4 cpus and lots of memory banks -- the cpus share the two mem and pci busses. It can take up to 32 1GB dimms, for a total of 32GB of ram. It's a Compaq ES40 Model II.
Because we run primarily single-process, single-threaded code, we have one or more users per cpu, instead of one or more cpus per user. This also saves administrative costs, because there is less hardware to deal with.
Slashdot Mirror
I don't think code reuse is an issue for TA. The original developers have split up. Furthermore, Cavedog, the company that made TA, was an experimental effort by some Humongous Entertainment (HE) employees, and operated under HE. HE primarily makes (some very good) children's games.
Who knows what bugs already existed when it was released (who knows that will tell us, anyway)?
Since it's not actively maintained, who knows what other bugs it might have (or expose in the current version of DirectX)?
Since nobody from Cavedog or HE would help look at the TA code, it seems unlikely that another company would want to mess with it at all.
If you were writing an modern adventure game, would you try to reuse code from Zork?
-Paul Komarek
The FFT doesn't count. It's math, not software. It's "just" a clever matrix factorization, not a huge software system. If you include the fft, you might as well go back to 3000BC and the decimal number system, or closer to 500AD and the discovery of how useful "zero" is.
Lapack is much closer to being impressive, but again it's "just" a math library. I'm a math grad student, and I appreciate the FFT and Lapack. But as software goes its longevity isn't as surprising as that of SABRE.
-Paul Komarek
I grew up in Spokane, WA, outside of Fairchild AFB. At the time, they had B-52 bombers there. I remember seeing a picture in the newspaper of about 8 people (2 per side, or something like that) in uniform carrying a large flat chunk of metal across a field. It was part of a flap that had broken off a B-52.
That "part of a flap" had about the same area (projected to 2d) as a Cadillac. Seems like the newspaper said that there were known flap problems at the time, and having flap parts fall off was a somewhat regular event. Thankfully Fairchild AFB is not particularly close to any population centers.
-Paul Komarek
Okay, I'm about to go really off the main topic. But about buying first generation equipment:
There is a small collection of old computer equipment behind some glass in a hallway of CMU's new computer science building. There is a picture of some long-forgotten computer being unloaded from a large truck, back in the early, early days (i.e. just after the Univac changed everything in the US computing industry). They give some advice to the reader, which they gleaned from : "Never buy serial number 1 of *anything*".
-Paul Komarek
The real problem here is figuring out what people mean by a "PC". I agree that the Xbox is not, in the strict sense, a PC. Then again, what is a PC these days? What is the most recent PC standard? The IBM PC standard wasn't even sacred for very long (i.e. once the BIOS was reverse engineered).
However, in the sense that "PC" means computer running windows (I really hate it when people do that), then I suppose the Xbox is mostly like a PC. If PC means being afflicted by legacy, then the Xbox's use of the x86 ISA places it firmly in the aflicted camp. If PC means endorsed by Microsoft, or compliant with whatever Microsoft comes up with today ("Designed for Microsoft Windows Whatever"), then it's a PC.
I guess the real issue on people's minds is how much work is necessary to port from the Xbox what we normally call an x86-based PC. I'm guessing that it would take a lot of work.
-Paul Komarek
I don't know if it's any more complex electrically, but certainly it is mechanically. The Xbox motherboard reminds me of the circuitboards I found as a kid when disassembling broken appliances from the scrapyard. Looking at it makes me want to get the soldering gun out and pull those valuable capacitors to put in my own projects!
-Paul Komarek
I agree with your assesment. The Gamecube strategy makes more sense when you look at it as a very powerful piece of "embedded" hardware designed to run (soft) real-time applications. It provides predictable, mostly on-time behavior, where the predictability comes from the simplicity. Thinking of the Gamecube as a weak general purpose computer is probably a mistake, or at least useless.
I think there's another point to be made here: Nintendo understands that fun games and "high-performance" are orthogonal. Think of how many fun games require a piece of cardboard with an array of alternatingly-colored squares. That cardboard has *really bad computational performance*, but that doesn't hurt it any. And besides, that cardboard is probably the most stable and lowest-power gaming platform around.
-Paul Komarek
You've just inspired another thought for how to use a GBA linked to a Gamecube. How about something like a murder mystery. Without private displays, this type of game would be practically impossible.
-Paul
Nintendo has always been somewhat enigmatic. Consider that the Gamecube is the first Nintendo console to abandon cartridges. I remember people suggesting that the N64 would die because of the limits cartridges imposed. Maybe Nintendo just understands the demographics better than Sony or Microsoft (or most of us) do, and in particular understands what makes things fun.
Mario Cart with four people in the same room has got to be way more fun than Mario Cart would be online. I know that Quake is way more fun when the room temp has soared to 85F (despite the window and box fans you've set up), you're bumping elbows with your opponents, and people answer when you shout "did you see how far my head flew?" (followed by someone using the toaster or microwave, blowing a fuse deep in the basement of your ancient house, thereby knocking all the computers and lights out).
-Paul Komarek
"and Zelda game out..."
Typo or Freudian slip? You decide.
-Paul Komarek
The Mozilla guys had a good explanation of why 1.0 is special, at least for them. 1.0 means API freeze -- in the sense that the 1.0 API will be available at least until 2.0 (though it could be extended). This is not so important for users, but for developers outside the Mozilla project, it is critical.
Since Evolution has a plugin facility, maybe the 1.0 obsession exists for similar reasons. Now developers can target it with peace of mind.
-Paul Komarek
Ticketmaster -- agreed! My knowledge of Microsoft networking comes mainly down to my experience, and various whitepapers at Carnegie Mellon University about why they disallow several Microsoft network services. Active Directory is a new thing for MS, and you're right that I'm ignorant on that account. But I have good reason to "think so little of Microsoft's accumen". They've stunk up networking for years.
As for being one of the most targetted networks, I'd agree. However, I'm not so sure they're number 2. I expect that other governments, Yahoo, and places like CERT (hosted by Carnegie Mellon) are also big targets.
As for Adam Smith, he *was* wrong. That's why the US (and every other 'capitalist' economy) uses a regulated implementation of capitalism. The free market makes its decisions based on marketing. If it made decisions based on research and development, then R&D would get 50% of revenues and marketing would get the 5% that R&D gets now.
Where is Microsoft today? Using hundreds of programmers to slowly reinvent unix networking. Why slowly? Because they're waiting for people to forget all the FUD they've put out about how bad unix networking is.
And I don't care about Microsoft's 2 million object directory. Why? Because I can piss farther than you!
-Paul Komarek
While most of what you say sounds reasonable, one thing really caught my eye: "only a few companies have people with enough experience with huge, varied networks". The problem with Microsoft is that they only have experience with huge, homogenous networks; they were blindsided by the internet; they thought remote admin was a bad idea until recently; their network hacks (netbios, for instance) stink on large networks.
I think Microsoft is very *unlikely* to have much useful exerperience with "huge, varied networks". What really gets me is that they seem to *like it this way*.
-Paul Komarek
On the other hand, if they manage to get the story down to 15 minutes, the remaining 75 minutes are really going to suck. And whose going to pay $8 for a 15 minute movie?
-Paul Komarek
I think people have forgotten just how bad the Metallica (was his name Lars or something?) phone interview went. The Metallica guy came off as a complete dick, ass, and idiot. I believe that interview was transcribed more-or-less exactly, and several people commented that Lars or whoever really did sound that dumb in "real life".
When people sit down and write, they're much more careful than when they talk. Oral and literate styles are very different. When someone answers the questions via email, they have a chance to reread the questions and edit their responses, possibly doing so multiple times. Yet some people still screw it up, like the Bush campaign in their slashdot interview.
If we've forgotten just how bad it can be, then we are definitely spoiled brats. I'm glad Bruce cared enough to make the phone call, and that Chris did the best he could at the time.
-Paul Komarek
While I agree that waiting a month wouldn't be so bad, I'd like to say thanks to Chris for giving it his best shot. So maybe he made a bad decision, but at least he tried to make the best of an apparently awkward situation.
I think the reason we wouldn't mind waiting a(nother) month is that we've gotten used to it. I simply forget about the interview once the questions are posted, and am later delighted to find the responses!
-Paul Komarek
I don't think they "disapprove of our way of life in the strongest way." I think they disapprove of our foreign policy. It's not to hard to extend that to disapproving of our collective apathy with respect to our government's foreign policy. All in all, I doubt they care much about us eating twinkies and watching sitcoms.
-Paul Komarek
Marx predicted the dominance of capitalism. I think it's Lenin you have a problem with. I believe Marx even wrote a letter to Lenin explaining why Lenin's perversion of Marx's writings were a bad idea. In fact, Marx's writings are probably more relevant than ever as we watch the battle against corporatism unfold (think "Battle of Seattle" and such).
-Paul Komarek
Whatever one might think of Richard Stallman, it is clear that he causes people to sit up and think. Every time there is an article about RMS on slashdot, an long debate follows. That's quite a contribution for one guy who probably doesn't read slashdot. He makes people think; obviously not everyone who posts is thinking, but at least a few are.
-Paul Komarek
Anyone else notice that L-wave's submission takes a sentence directly from The Register's John Lettice? Look at http://www.theregister.co.uk/content/4/22928.html.
-Paul Komarek
Even if you do use original packaging including molded styrofoam, those clever UPS folks will find a way to damage your shipment. While the damage is likely to be less than what this guy photographed, it's still damage.
We received a rackmount APC Uninteruptable Power Supply (overloaded abbreviations!) via UPS ground. The front plate is steel, about 5mm (about 3/16") thick. It was severaly bent, and we had to straighten it with a benchtop vice.
I've shipped a computer in the original packaging via UPS ground. When it arrived, some case screws had popped out, the case frame was bent, and the plastic front had broken pieces. Needless to say the case didn't really fit right after that.
I've never had damage with FedEx ground, though I've had two computers shipped that way (in original packaging). I had to pick one up from the FedEx distribution location because they messed up the delivery (never put it on the truck), and the box was banged up a little, but the case was fine.
Whenever a vendor doesn't offer FedEx ground, I write a note requesting they consider it.
-Paul Komarek
You're fine. Never use an apostrophe with a possessive pronoun like "its". The original guy is whacked. I've never been completely comfortable with "it's" for "it has" in written language. I'm really uncomfortable with "who's" in written language. Obviously both are spoken.
-Paul Komarek
Although the thesis sometimes feels like it takes 500 years...
-Paul Komarek
I'll put in a plug for Python. It's clean, cross-platform, easy to learn (a core focus of the language is making things easy for nonprogrammers), and has support for QT, GTK, Tk, and some other widget libs. There are good books available for learning Python, and the Python community is fiarly strong.
Python is interpreted, so you'll won't see a lot of speed from it. Most applications don't actually need much speed, especially gui apps. However, you wouldn't want to built the SETI-at-home back end in Python, for example.
-Paul Komarek
Microway's prices on the website are horribly out of date. At the beginning of this year, we bought two dual 264DP machines with 4GB each in rackmount cases with slide for $13,500 each. I'm sure the "loaded" config is less now.
That's still expensive, but if you need a *single* *fast* cpu, a bunch of dual P3-1GHz won't do it for you. We need the large virtual memory, and more than that, our code is single-process and single-threaded. We just aren't into clustering, primarily for historical reasons (large, old codebase among others). Other labs might do things differently -- deal with their own memory allocators to span processes, and handle the extreme NUMA-ishness of a cluster. We'd rather put our money up front and save time.
We've got a bigger machine which is basically a 264DP with 4 cpus and lots of memory banks -- the cpus share the two mem and pci busses. It can take up to 32 1GB dimms, for a total of 32GB of ram. It's a Compaq ES40 Model II.
Because we run primarily single-process, single-threaded code, we have one or more users per cpu, instead of one or more cpus per user. This also saves administrative costs, because there is less hardware to deal with.
-Paul Komarek