Heh. This is probably as silly as it gets, answering old posts that noone will ever read, ever.;^) Anyway, I think I'm entitled to have issues with something from just reading or hearing about it, i.e., without actually trying it. Of course you may disagree; that's just the way I am. Also, I thought I was defending Python here, not doing some kind of attack against its choice of block syntax. I'm sorry for trying.
I'm no real expert in these matters, but I would go as far as to say that any company having the skillz to reverse engineer and manufacture a clone board, surely should have the ability to disassemble the driver. Sure, giving them it in source form from the start might lessen the amount of work required for the reverse engineering, but I don't think it prevents it much. I'm thinking a bit about NVIDIA here, since I find their approach interesting: a partially open-sourced kernel driver, which interfaces with a binary-only userspace driver. Maybe that could work for the AnonymousIntern's company's products?
Um, I've never used Python (I have issues with the block syntax), but I think most users of it would seriously protest against your classification of Python as a "web scripting language". In fact, by reading the What is Python? page over at www.python.org, I get the impression that Python is "an interpreted, interactive, object-oriented programming language". See? Nothing about it being specifically for the web there...
Never do math (even with the help of the excellent grpn calculator) late in the afternoon. Since every frame is 20 MB, 9 MB is clearly not even enough for a single frame, so never mind the mention of double buffering above. Thanks.
Thinking about the same thing, I did some simple calculation on bandwidth requirements... 2560x2048 is (as stated) 5Mpixels (using M to mean 1<<20, as in MB). At 32 bits per pixel, that's 20 MB per frame. If we want to display that at 60 Hz, that's a rather hefty 1.2 GB per second bandwidth requirement. One way to ease that is to split the frame buffer across multiple cards, since each frame buffer then only needs to deal with a fraction (here, a fourth, or 300 MBps) of the bandwidth. Reservation: 60 Hz might be more than an LCD uses, so the above figures could be off by a factor of 2 or so. Still, I think there's a problem here.
If you want to do full-screen 3D graphics (which seems to be high on everybody's wish list, judging from the number of drooling references to Q3 among the posts here:), you do not want your display to "steal" 1.2 GBps of the available bandwidth. You want to use that bandwidth to blast pixels to the screen, read textures and Z-buffer values, etc. It's unclear what the solution to this might be. You could go the route of Bitboys, and embed the frame buffer in the display core, thus giving wide busses and huge (they mention 12 GBps) bandwidth. However, embedded DRAM currently limits the amount of memory in the frame buffer rather strictly. Bitboys talk about 9 MB, which in the case of the Roentgen wouldn't even allow double buffering the entire screen (remember, each screenful requires 5 MB)... My point, then? Um, I don't know. Memory bandwidth is a hairy thing, or something. Good luck to all engineers involved!;^)
Um, not wanting to sound like some GPL-fanatic, but is this really a problem? If "they", i.e. a bunch of the world's largest hi-tech companies, make a mistake which legally forces them to deliver source code with all their cool products, who loses? Certainly not "us", i.e. the geeks (who also happen to be consumers) of the world. Or? If they didn't want to have to deliver source, one would hope that they would be smart enough not to pick a kernel licensed under the GPL, right? I think I don't quite see what's worth sighing about here. Unless, of course, you're afraid that they plan to violate the GPL by withholding the source. That could be a problem. FSF's lawyers against that consortium's could become bloody...:(
Heh. I hope you (and everybody else) realise that the blue LED is probably there just because the PlayStation 2 happens so have one. ObAnecdote: I had the pleasure of listening in to Phil Harrison's (VP of SCEA) key note at the GDC in March, and he made a point of the fact that the blue LED on the machine seems to be the most successful part of its design.;^) Apparantly, people couldn't care less for Emotion Engines, Graphics Synthesizers, and analog controller buttons -- just give them a nice blue LED to look at and they're happy!;^)
Um, pardon my French, but who the fsck died and made them all gods of the Internet or something? Can someone please define the concept of a "rogue" website? As opposed to what, a nice corporation-controlled e-commerce site? And what do they mean by saying that they can "remove offending messages" from the net? Geez. I think we all need another look at both Fling and FreeNet. This sucks. I'm almost upset.
Um, I don't know, but they're building a database, perhaps mapping IP pairs to network distance (hops and/or ping time), right? What can be done with such a database, then? Um... I don't know. How about use the database, in combination with some clever software, to optimize routing tables globally? That might be packagable into a sellable service... Or it might be totally impossible. Ack, too long since I took my internetworking courses. Shoot me down, network geeks!;^)
Heh. Reminds me of a day at work a couple of months ago, when a colleagues' box was hacked into. The h4xx0r kid had run some kind of rootkit (although I'm not sure the box was actually rooted, but some kind of prepackaged kit was used), which cleaned out all the logs. Except, of course, for that tricky, well-hidden, hard-to-find, sneaky, known-by-gurus-only one known as.bash_history!;^)
It was quite cool to see which commands had been run, etc. I think he actually started up an IRC server on the box, probably to serve warez... That, and the ObPortscan of course.;^) Our local nettech archived the contents of the kid's account down on a couple of floppies and did a reinstall. Some day, when I'm sufficiently bored, I think I'll ask him for those disks. Might learn something useful.;^)
Um, I'm sure you're aware of this, but being able to translate an application's strings without access to the source is not heavy magic. It is, for instance, a standard feature of GNU gettext, the library most often used when translating Linux (and other open source) applications. And since modern toolkits (such as GTK+) doesn't do layout based on pixel coordinates, reflowing is automatic.
Ouch! That reminds me of a current ad campaign visible on the streets of Stockholm. One of the messages is "You have a sweaty broker in your pocket".;^) (I think the campaign is for WAP phones/services, or something)
Um, I would like to have a look under the hood of your PC sometime. 200 MHz DDR-SDRAM on a 128-bit bus is nowhere to be seen in mine.;^( BY the time the Xbox actually ships, though, I hope that will have been resolved... Anyway, what I'm trying to say is that DDR-SDRAM is only found on graphics cards today, not as main (system) memory.
The Xbox is specified to include a custom (=proprietary) video output module, with the possibility connecting the mythical HDTVs of next year. Or next... Anyway, the specified maximum resolution supported is 1920x1080 pixels.
No, because the Apple Macintosh, for hysterical raisins, lacks memory protection. That, and the fact that their mouse designers can't count to three, makes the platform both unstable and generally unusable. </RANT>
The PS2 also has PC Card and FireWirre (IEEE1394), of course. The Xbox uses USB for its controllers, only (since it is Microsoft), they don't use the physical part of the standard, "only" the electric and signalling. I.e., the connectors won't look like USB, so Microsoft can easier control licensing to manufacturers, or something. Of course, I think that might be circumvented pretty easily...
I don't know about "realistically", but SGI have a few boxes meant for the desktop, that run Irix, a 64-bit operating system. The O2, for example, or the (newly announced, I think) Octane2... Of course, these machines are not within most people's realistic budgets, but they do exist...;^)
My biggest gripe with Mozilla, from using the milestones a little, is the slowness of the GUI. I understand that the interface is implemented in some super-dynamic fashion internally (based on XML), but I still think it's too slow, and also less visually attractive than e.g. a plain GTK+ UI. They used to have a stripped-down version of the browser which I think used GTK+ natively, but it seems to have been dropped. Presumably, it is possible to reimplement the GUI using existing mechanisms, but I don't know if anyone has given that a try yet. I'm quite fluent with the keyboard, and having to wait the split second it takes for Mozilla to open the "Find in This Page..." dialog really gets to me (especially on an Athlon 550 MHz machine).
Just wanted to say to the author, that this might be one of the better pieces I've read on Slashdot lately. Shock-full of neat technical info, great links, and just felt solid and correct! Thanks you very much!
Of course I call strcpy(). That's an even more powerful idiom, with the added benefit of being standardized, so that it can be incredibly cleverly implemented/optimized if the vendor felt like it. Still, I don't think that particular while loop is that horrible...
Um, since it's ESR we're talking about, I do hope you mean that in some kind of metaphorical sense, since the literal interpretation threatens to make the community one spokesperson poorer.;^)
He. If you think you know C, you should recognize that strcpy() implementation as a verbose version of K&R2 page 106. If you don't, you probably haven't read K&R, and it then immediately follows that you don't know C.;^) Also, I think your idea that all code should be commented sort of goes against the idea of idioms in languages. The above snippet is an idiom, and an experienced C programmer will recognize it immediately as such. Of course, someone with less experience wouldn't, but then it becomes a learning experience!
Hey, that's interesting! I grew up hacking on the Amiga, but mainly learned C over in Linux-land... I never did much programming in AmigaOS, but even when I did, I don't think I knew enough C for it to seem weird, something your snippet above does today.;^) The thing is, the Amiga's handling of libraries (.so:s, or DLL:s) is weird (from a C perspective). You open (load) a library by calling OpenLibrary(), and get a pointer known as the library base back. In assembly, you would then jump into a function table at an offset from that pointer. In C, you don't use the pointer, you just magically call the library function directly, relying on code in stubs to use the library base. This doesn't seem very clean today, really... It would be cleaner if you actually used the library base as an explicit function table, e.g. IntuitionBase->OpenWindow(...) and so on. Hm. OK, just a thought, move along now people.
Slashdot Mirror
Heh. This is probably as silly as it gets, answering old posts that noone will ever read, ever. ;^) Anyway, I think I'm entitled to have issues with something from just reading or hearing about it, i.e., without actually trying it. Of course you may disagree; that's just the way I am. Also, I thought I was defending Python here, not doing some kind of attack against its choice of block syntax. I'm sorry for trying.
I'm no real expert in these matters, but I would go as far as to say that any company having the skillz to reverse engineer and manufacture a clone board, surely should have the ability to disassemble the driver. Sure, giving them it in source form from the start might lessen the amount of work required for the reverse engineering, but I don't think it prevents it much. I'm thinking a bit about NVIDIA here, since I find their approach interesting: a partially open-sourced kernel driver, which interfaces with a binary-only userspace driver. Maybe that could work for the AnonymousIntern's company's products?
Hey, that reminds me of something I read in the Jargon File: namely ICBM addresses!. ;^) Feels dangerous...
Um, I've never used Python (I have issues with the block syntax), but I think most users of it would seriously protest against your classification of Python as a "web scripting language". In fact, by reading the What is Python? page over at www.python.org, I get the impression that Python is "an interpreted, interactive, object-oriented programming language". See? Nothing about it being specifically for the web there...
Never do math (even with the help of the excellent grpn calculator) late in the afternoon. Since every frame is 20 MB, 9 MB is clearly not even enough for a single frame, so never mind the mention of double buffering above. Thanks.
Thinking about the same thing, I did some simple calculation on bandwidth requirements... 2560x2048 is (as stated) 5Mpixels (using M to mean 1<<20, as in MB). At 32 bits per pixel, that's 20 MB per frame. If we want to display that at 60 Hz, that's a rather hefty 1.2 GB per second bandwidth requirement. One way to ease that is to split the frame buffer across multiple cards, since each frame buffer then only needs to deal with a fraction (here, a fourth, or 300 MBps) of the bandwidth. Reservation: 60 Hz might be more than an LCD uses, so the above figures could be off by a factor of 2 or so. Still, I think there's a problem here.
If you want to do full-screen 3D graphics (which seems to be high on everybody's wish list, judging from the number of drooling references to Q3 among the posts hereUm, not wanting to sound like some GPL-fanatic, but is this really a problem? If "they", i.e. a bunch of the world's largest hi-tech companies, make a mistake which legally forces them to deliver source code with all their cool products, who loses? Certainly not "us", i.e. the geeks (who also happen to be consumers) of the world. Or? If they didn't want to have to deliver source, one would hope that they would be smart enough not to pick a kernel licensed under the GPL, right? I think I don't quite see what's worth sighing about here. Unless, of course, you're afraid that they plan to violate the GPL by withholding the source. That could be a problem. FSF's lawyers against that consortium's could become bloody... :(
Heh. I hope you (and everybody else) realise that the blue LED is probably there just because the PlayStation 2 happens so have one. ObAnecdote: I had the pleasure of listening in to Phil Harrison's (VP of SCEA) key note at the GDC in March, and he made a point of the fact that the blue LED on the machine seems to be the most successful part of its design. ;^) Apparantly, people couldn't care less for Emotion Engines, Graphics Synthesizers, and analog controller buttons -- just give them a nice blue LED to look at and they're happy! ;^)
Um, pardon my French, but who the fsck died and made them all gods of the Internet or something? Can someone please define the concept of a "rogue" website? As opposed to what, a nice corporation-controlled e-commerce site? And what do they mean by saying that they can "remove offending messages" from the net? Geez. I think we all need another look at both Fling and FreeNet. This sucks. I'm almost upset.
Um, I don't know, but they're building a database, perhaps mapping IP pairs to network distance (hops and/or ping time), right? What can be done with such a database, then? Um... I don't know. How about use the database, in combination with some clever software, to optimize routing tables globally? That might be packagable into a sellable service... Or it might be totally impossible. Ack, too long since I took my internetworking courses. Shoot me down, network geeks! ;^)
By reading the article, I learned that those are actually space keys.
Heh. Reminds me of a day at work a couple of months ago, when a colleagues' box was hacked into. The h4xx0r kid had run some kind of rootkit (although I'm not sure the box was actually rooted, but some kind of prepackaged kit was used), which cleaned out all the logs. Except, of course, for that tricky, well-hidden, hard-to-find, sneaky, known-by-gurus-only one known as .bash_history! ;^)
It was quite cool to see which commands had been run, etc. I think he actually started up an IRC server on the box, probably to serve warez... That, and the ObPortscan of course.Um, I'm sure you're aware of this, but being able to translate an application's strings without access to the source is not heavy magic. It is, for instance, a standard feature of GNU gettext, the library most often used when translating Linux (and other open source) applications. And since modern toolkits (such as GTK+) doesn't do layout based on pixel coordinates, reflowing is automatic.
Ouch! That reminds me of a current ad campaign visible on the streets of Stockholm. One of the messages is "You have a sweaty broker in your pocket". ;^) (I think the campaign is for WAP phones/services, or something)
Um, I would like to have a look under the hood of your PC sometime. 200 MHz DDR-SDRAM on a 128-bit bus is nowhere to be seen in mine. ;^( BY the time the Xbox actually ships, though, I hope that will have been resolved... Anyway, what I'm trying to say is that DDR-SDRAM is only found on graphics cards today, not as main (system) memory.
The Xbox is specified to include a custom (=proprietary) video output module, with the possibility connecting the mythical HDTVs of next year. Or next... Anyway, the specified maximum resolution supported is 1920x1080 pixels.
No, because the Apple Macintosh, for hysterical raisins, lacks memory protection. That, and the fact that their mouse designers can't count to three, makes the platform both unstable and generally unusable.
</RANT>
The PS2 also has PC Card and FireWirre (IEEE1394), of course. The Xbox uses USB for its controllers, only (since it is Microsoft), they don't use the physical part of the standard, "only" the electric and signalling. I.e., the connectors won't look like USB, so Microsoft can easier control licensing to manufacturers, or something. Of course, I think that might be circumvented pretty easily...
I don't know about "realistically", but SGI have a few boxes meant for the desktop, that run Irix, a 64-bit operating system. The O2, for example, or the (newly announced, I think) Octane2... Of course, these machines are not within most people's realistic budgets, but they do exist... ;^)
My biggest gripe with Mozilla, from using the milestones a little, is the slowness of the GUI. I understand that the interface is implemented in some super-dynamic fashion internally (based on XML), but I still think it's too slow, and also less visually attractive than e.g. a plain GTK+ UI. They used to have a stripped-down version of the browser which I think used GTK+ natively, but it seems to have been dropped. Presumably, it is possible to reimplement the GUI using existing mechanisms, but I don't know if anyone has given that a try yet. I'm quite fluent with the keyboard, and having to wait the split second it takes for Mozilla to open the "Find in This Page..." dialog really gets to me (especially on an Athlon 550 MHz machine).
Just wanted to say to the author, that this might be one of the better pieces I've read on Slashdot lately. Shock-full of neat technical info, great links, and just felt solid and correct! Thanks you very much!
Of course I call strcpy(). That's an even more powerful idiom, with the added benefit of being standardized, so that it can be incredibly cleverly implemented/optimized if the vendor felt like it. Still, I don't think that particular while loop is that horrible...
Um, since it's ESR we're talking about, I do hope you mean that in some kind of metaphorical sense, since the literal interpretation threatens to make the community one spokesperson poorer. ;^)
He. If you think you know C, you should recognize that strcpy() implementation as a verbose version of K&R2 page 106. If you don't, you probably haven't read K&R, and it then immediately follows that you don't know C. ;^) Also, I think your idea that all code should be commented sort of goes against the idea of idioms in languages. The above snippet is an idiom, and an experienced C programmer will recognize it immediately as such. Of course, someone with less experience wouldn't, but then it becomes a learning experience!
Hey, that's interesting! I grew up hacking on the Amiga, but mainly learned C over in Linux-land... I never did much programming in AmigaOS, but even when I did, I don't think I knew enough C for it to seem weird, something your snippet above does today. ;^) The thing is, the Amiga's handling of libraries (.so:s, or DLL:s) is weird (from a C perspective). You open (load) a library by calling OpenLibrary(), and get a pointer known as the library base back. In assembly, you would then jump into a function table at an offset from that pointer. In C, you don't use the pointer, you just magically call the library function directly, relying on code in stubs to use the library base. This doesn't seem very clean today, really... It would be cleaner if you actually used the library base as an explicit function table, e.g. IntuitionBase->OpenWindow(...) and so on. Hm. OK, just a thought, move along now people.