The mods must be smoking crack today. This should be +5 Insightful, not 0 Troll. It's pretty much dead on.
I would also note that the current shuttles are not first rev hardware. The Columbia was the last of the first generation. Designed to schematics that predated those of Challenger, it had the disadvantage of being both the smallest and heaviest shuttle in the fleet. It blew up in large part because of that. Had the Atlantis or Discovery been in its place, I think there's a very good chance that the reduced descent heating would have been enough for it to get down in one piece, or at least to get far enough down for emergency egress to be possible.
A lot of people said that Columbia should have been retired when they announced their plans to retrofit it. They were right. It should never have been retrofitted with new electronics. It should have been put in a museum or scrapped. It was an old fossil with plenty of design mistakes.
But the current shuttles are third-generation tech, IIRC. The Challenger was significantly upgraded from the first rev, and the ones that are still intact are all significantly larger, lighter, and have larger payload capacity. Even in the third generation, though, all the things the parent poster mentioned still haven't been fixed, making the shuttle launches obscenely expensive relative to their payload capacity and general scientific usefulness.
Frankly, this is the best news I've heard in a long time. The ESA and the Russians should be able to learn from the mistakes of the shuttle and get things right, assuming they can get past all the politics and corner-cutting. It's like I've always said: build it right to begin with and you won't have to keep rebuilding it....
When a CEO bitches like that, he's just scared of competition.
Terrified is more like it. With one of the biggest infrastructures and the largest customer base of any cell provider in the U.S., they have the most to lose if their overpriced, unreliable (IMHO) service gets encroached upon by much cheaper and only slightly more unreliable services.
I'm willing to bet drive manufacturers -do- have custom firmwares that do that. Why? Because otherwise they would end up generating a lot of bricks while testing bug fixes to those parts of the firmware....
Further, it shouldn't be that hard to solve this problem. The drive reads the data off the disk. There's a ribbon cable between the controller board and the disk. Tap the data stream. Feed it into a logic analyzer that has a digital data ouptut (e.g. a USB logic analyzer). Take the data captured, find the sync bytes, then shove the remainder into an RLL decoder.
Now figure out the ECC format used (it will typically be four bytes at the end of each sector, but this may vary). Strip the ECC bytes. You now have a track image of the track in question, probably with some extra sync bytes between sectors, but I'm not sure. If you want, you could simply single-step the drive motor repeatedly and copy the entire disk this way, but it is probably more effective to write a program that scans for things that right be an ATA password and tries them sequentially.
To make this easier, every 4 passwords or so, the tool should ask you to power-cycle the drive. To facilitate this, take a power extender cable and cut the 5v line. Put a momentary off pushbutton inline. Press for a second and then release. In all likelihood, you should only need to power cycle the drive electronics, not the drive motor (12v).
I've never tried this, of course, but in principle, it shouldn't be that bad....
Yeah, the McCarthy hearings pretty much worked the same way. I find it particularly telling that the McCarthy hearing transcripts were required reading for the Senate's Homeland Security and Governmental Affairs Committee recently.... See this page.
Those who cannot learn from history are doomed to repeat it.
---George Santayana
That's what concerns me. We put the Hubble up there and immediately found critical problems that made it nearly useless without corrective optics. What would prevent us from screwing up the replacement? And if they put it at the LaGrange point, we're doubly screwed.
If history has taught us anything, it is that the replacement is only cheaper if it works perfectly the first time. I suspect the cost estimates are based on current test practices which are insufficient for ensuring that it will work perfectly the first time, as we have repeatedly proven through screw-ups in the past. Thus, the probability leans towards the costs being far higher than estimated, whether as a result of doing extra testing or as a result of going back and fixing the mistakes later.
Of course, the worst case scenario would involve trying to figure out a way to get a shuttle to the LaGrange point (which I'm told is impossible without significant modifications to the current shuttle).
If I believed for a single second that they could replace the Hubble with a new one that worked correctly for less than the cost of repairing it, I'd be shouting "dump it" as fast as the next guy, but I'm far too cynical to do anything more than laugh at the notion.
The sad thing is that this problem is easily solved with tools. For example, a lot of Apple's man pages are now being created with the MPGL (man page generation language). It's a little open source tool that turns an HTML-like language into mdoc man pages.
Shouldn't be particularly hard to adapt Texinfo to spew something like this. Texi2latex converts Tex to Latex via XML. Just stop it halfway and add some code to transform the XML dtd. Better, modify texi2html to push out a thinned html with appropriate headings.
Umm... no. RISC wins most of the time if all other things are equal. If that weren't so, you'd be using a CISC chip right now. What's that? You think your AMD or Intel chip is CISC? Not if it's recent. It's a RISC core. Those CISC instructions are broken into micro-ops that are then processed in a RISC fashion.
Why is that, you ask? Because RISC instructions are inherently more pipelineable. The only place -CISC- wins is if all the instructions must be sequential. RISC can do far more instruction reordering such as prefetching a chunk of data significantly farther ahead of its use because that fetch instruction is a separate, schedulable entity. Thus, to give CISC equivalent performance, it was necessary to split CISC instructions into roughly RISC-equivalent components. This is done in pretty much every modern CISC chip out there.
As for CISC compilers optimizing the code better, that's probably true. Now imagine what would happen if the code you were running on a RISC box were as well optimized.... Suddenly the RISC machine isn't just keeping up with the CISC. It is leaving it in the dust. Try IBM's POWER/PowerPC compilers some time....:-)
And all that extra hardware for cracking pretty much every CISC instruction into multiple instructions? Extra heat, extra die space that could be used for cache, extra CPU cycles to do the cracking, extra power consumption....
Basically the only way CISC wins is if you stack the deck.
As a general rule, most stores say that if their inventory shows less than three of something, the real number is very likely zero. This can happen because of shoplifting, but more commonly is caused by human error---when the scanner doesn't pick up the number and the person keys it in manually and punches in the wrong number (rarely if you're using UPCs, but common with store-generated numbers that often lack check digits), when someone miscounts the number of items on the shelf during inventory, when a cashier sees two similar items and rings them up by scanning one twice, etc.
*sigh*. The whole idiot-proof thing and underestimating the ingenuity and/or persistence of idiots comes to mind.
And they probably added the cost of Mac IE to the cost of Mac Office... which explains why its professional edition costs as much as the Mac Mini....
I think we should start describing software cost in terms of the amount of hardware you can buy for the price. This should put things into better perspective. Office: Mac Mini. Photoshop: edu price for an eMac. Saying "screw it, I'm downloading OO and the GIMP": priceless.
The problem is that right-click and control-click are reserved for contextual menus. There's a Technical Q&A on the subject. I'm -very- surprised that control-dragging works. It should exhibit the same behavior.
You should consider using middle-drag (option-drag) on the Mac, since that should not exhibit this problem.
TI might have been involved in ADB; I'm not sure. You're probably thinking of NuBus, though, which TI originally designed and Apple used for a long time. Apple and TI have collaborated a good bit over the years, working together on FireWire, for example.
However, you're actually wrong about it being only used by Apple. NeXT also used ADB. While NeXT is part of Apple now, it wasn't at the time....
A little trivia.... The first Apple computer with ADB was the Apple IIgs. It predated its introduction on the mac (in the Mac SE) by a year or so.
At the time, Apple probably could have used PS/2, but instead used their own standard that was far more flexible, and generally a much more intelligent bus. The ADB bus could support... I think 16 peripherals on the bus, complete with a device ID resolution mechanism, hot swappability, etc. Basically, it did most of what USB did, but a decade earlier, albeit at lower bandwidth and with a decidedly more linear signal path.;-)
Very nice little bus for its time.... Way, way ahead of its time.... As usual, Apple creates and/or chooses a superior technology and everybody else chooses a less flexible, more CPU-intensive bus because it's cheaper. *Sigh*.
Even when the hardware is doing some of the work, it's usually not doing all of it, a few specialized MPEG-to-video output cards notwithstanding. There are several different parts of playing back MPEG video, including: decompressing, deinterlacing (optional), scaling, color space conversion, etc. The hardware only provides the last one, and sometimes the next-to-last, IIRC. The actual decompression occurs in software (doing all the inter-frame math, etc.), as does deinterlacing, generally, assuming you are using an MPEG player that supports it.
Basically, doing full MPEG decode in the GPU made sense before we had vector units in the CPU, but these days, it isn't enough of an advantage to care. And besides, you usually want to have some software control over the presentation---the ability to float windows over it, etc. That sort of thing doesn't work too well if the GPU is doing all of the decoding for you.
And, of course, HD video is a different animal---different resolution, higher data rate, etc. Even if you have custom hardware to decode SD video and send it to a TV on-the-fly, doing the same for HD is sufficiently harder that it probably isn't worth it, particularly since you have a fast CPU handy already.
Which is exactly why I said HTML is for viewing , PDF is for printing. I did say that my rant was against people who only posted PDF without accompanying HTML, not against people who post PDFs for printing. Just don't use -only- PDF.:-)
PDF? Proprietary? Only if you mean Adobe's implementation. There are thousands of tools out there for generating and viewing PDF content in the open source world. Calling PDF proprietary simply because Adobe doesn't provide a viewer for all platforms would be like calling multicast DNS proprietary because at least initially, stock versions of Rendezvous wouldn't compile under Linux.
Based on that same definition, Postscript is proprietary. Oddly enough, Ghostscript is sometimes known to open encapsulated postscript files generated by Adobe Illustrator that Adobe's own Photoshop can't. When the open source software exceeds the quality and reliability of the reference implementation, it can no longer reasonably be described as proprietary, even if the reference implementation happens to be, IMHO.
That said, I would no more recommend people posting PDF or OOo docs than Word docs, for exactly the same reason. You have to download special software to view it. Even if Firefox had a plug-in in the shipping version, most people wouldn't have that version. For that matter, most people don't use Firefox.
The web is a powerful platform for deployment of information precisely because there are a very limited number of standard formats for contents, and a single standard environment for viewing them. It pisses me off to no end when I see a PDF file without an HTML version alongside it. The last thing I want to do is deal with a whole different environment to view content---whether it's Acrobat or a viewer plug-in makes no difference. Ditto for Word, OOo, etc. (As I always say, "Repeat after me: 'HTML is for Viewing, PDF is for Printing'.")
And I hope I -never- have to read something that some clueless peson uploaded in Postscript again. Yes, there's software for every platform, but no, most people don't have it installed, and it's a pain in the ass to distill to PDF just to view something that's usually mostly plain text anyway. And before you ask, yes, sometimes I have been known to just read the Postscript file in vi.
Bottom line, if in doubt, HTML. If HTML won't work because the person posting it is too anal about formatting... HTML anyway, and post a nice, neat, formatted PDF for the three other people in the world who are as anal as they are.;-)
</rant>
We now return you to your regularly scheduled discussion of open formats.
It's a little different there, actually. It's more like selling drugs to some two-bit thugs who then report you to the cops for selling drugs. If you sell to the cops, they are legally allowed to execute sting operations. It isn't always clear about private entities' rights to do the same.
That said, it's probably a fool's bet if you actually tried to use that defense.:-)
The disturbing thing is that CNN ran an article on this a few minutes ago and appears to have pulled it since, but it was so slanted it wasn't even funny. They even railed on BitTorrent (which is -heavily- used for non-infringing purposes).
I would strongly encourage everyone reading this to watch CNN's website and if the story reappears and is as slanted as the one I read, write to them and complain.
Umm... cancer is caused by genetic mutation, which is, by its nature, a driving force of evolution. Disease drives evolution by weeding out weaker members. Mutation drives evolution by creating stronger members, except that sometimes it doesn't work.
Thus, by stopping cancer, you're actually -helping- evolution by ensuring that only the useful and non-fatal mutations continue into the next generation.
Happily, they haven't updated the series 1 TiVo software in about two years. I'm guessing they probably won't for this, either. If you really want a TiVo that's free of this disgusting hack, buy a used Series 1 and upgrade the hard drive.
Slashdot Mirror
I would also note that the current shuttles are not first rev hardware. The Columbia was the last of the first generation. Designed to schematics that predated those of Challenger, it had the disadvantage of being both the smallest and heaviest shuttle in the fleet. It blew up in large part because of that. Had the Atlantis or Discovery been in its place, I think there's a very good chance that the reduced descent heating would have been enough for it to get down in one piece, or at least to get far enough down for emergency egress to be possible.
A lot of people said that Columbia should have been retired when they announced their plans to retrofit it. They were right. It should never have been retrofitted with new electronics. It should have been put in a museum or scrapped. It was an old fossil with plenty of design mistakes.
But the current shuttles are third-generation tech, IIRC. The Challenger was significantly upgraded from the first rev, and the ones that are still intact are all significantly larger, lighter, and have larger payload capacity. Even in the third generation, though, all the things the parent poster mentioned still haven't been fixed, making the shuttle launches obscenely expensive relative to their payload capacity and general scientific usefulness.
Frankly, this is the best news I've heard in a long time. The ESA and the Russians should be able to learn from the mistakes of the shuttle and get things right, assuming they can get past all the politics and corner-cutting. It's like I've always said: build it right to begin with and you won't have to keep rebuilding it....
Terrified is more like it. With one of the biggest infrastructures and the largest customer base of any cell provider in the U.S., they have the most to lose if their overpriced, unreliable (IMHO) service gets encroached upon by much cheaper and only slightly more unreliable services.
Oh, you mean -supported- city-wide WiFi....
Associated Press ran story that Tony Blair was courting Margaret Thatcher
I'm sure there are plenty more where those came from.
Send your rebate via registered mail. And if that doesn't work, you're screwed.
No, lesson learned: it costs 5 cents to xerox a receipt/rebate form and $100 not to.
Here, here! (sic).
Further, it shouldn't be that hard to solve this problem. The drive reads the data off the disk. There's a ribbon cable between the controller board and the disk. Tap the data stream. Feed it into a logic analyzer that has a digital data ouptut (e.g. a USB logic analyzer). Take the data captured, find the sync bytes, then shove the remainder into an RLL decoder.
Now figure out the ECC format used (it will typically be four bytes at the end of each sector, but this may vary). Strip the ECC bytes. You now have a track image of the track in question, probably with some extra sync bytes between sectors, but I'm not sure. If you want, you could simply single-step the drive motor repeatedly and copy the entire disk this way, but it is probably more effective to write a program that scans for things that right be an ATA password and tries them sequentially.
To make this easier, every 4 passwords or so, the tool should ask you to power-cycle the drive. To facilitate this, take a power extender cable and cut the 5v line. Put a momentary off pushbutton inline. Press for a second and then release. In all likelihood, you should only need to power cycle the drive electronics, not the drive motor (12v).
I've never tried this, of course, but in principle, it shouldn't be that bad....
Those who cannot learn from history are doomed to repeat it.
---George Santayana
If history has taught us anything, it is that the replacement is only cheaper if it works perfectly the first time. I suspect the cost estimates are based on current test practices which are insufficient for ensuring that it will work perfectly the first time, as we have repeatedly proven through screw-ups in the past. Thus, the probability leans towards the costs being far higher than estimated, whether as a result of doing extra testing or as a result of going back and fixing the mistakes later.
Of course, the worst case scenario would involve trying to figure out a way to get a shuttle to the LaGrange point (which I'm told is impossible without significant modifications to the current shuttle).
If I believed for a single second that they could replace the Hubble with a new one that worked correctly for less than the cost of repairing it, I'd be shouting "dump it" as fast as the next guy, but I'm far too cynical to do anything more than laugh at the notion.
See HeaderDoc User's Guide: Using the MPGL Suite
Shouldn't be particularly hard to adapt Texinfo to spew something like this. Texi2latex converts Tex to Latex via XML. Just stop it halfway and add some code to transform the XML dtd. Better, modify texi2html to push out a thinned html with appropriate headings.
Why is that, you ask? Because RISC instructions are inherently more pipelineable. The only place -CISC- wins is if all the instructions must be sequential. RISC can do far more instruction reordering such as prefetching a chunk of data significantly farther ahead of its use because that fetch instruction is a separate, schedulable entity. Thus, to give CISC equivalent performance, it was necessary to split CISC instructions into roughly RISC-equivalent components. This is done in pretty much every modern CISC chip out there.
As for CISC compilers optimizing the code better, that's probably true. Now imagine what would happen if the code you were running on a RISC box were as well optimized.... Suddenly the RISC machine isn't just keeping up with the CISC. It is leaving it in the dust. Try IBM's POWER/PowerPC compilers some time.... :-)
And all that extra hardware for cracking pretty much every CISC instruction into multiple instructions? Extra heat, extra die space that could be used for cache, extra CPU cycles to do the cracking, extra power consumption....
Basically the only way CISC wins is if you stack the deck.
*sigh*. The whole idiot-proof thing and underestimating the ingenuity and/or persistence of idiots comes to mind.
I think we should start describing software cost in terms of the amount of hardware you can buy for the price. This should put things into better perspective. Office: Mac Mini. Photoshop: edu price for an eMac. Saying "screw it, I'm downloading OO and the GIMP": priceless.
You should consider using middle-drag (option-drag) on the Mac, since that should not exhibit this problem.
However, you're actually wrong about it being only used by Apple. NeXT also used ADB. While NeXT is part of Apple now, it wasn't at the time....
A little trivia.... The first Apple computer with ADB was the Apple IIgs. It predated its introduction on the mac (in the Mac SE) by a year or so.
At the time, Apple probably could have used PS/2, but instead used their own standard that was far more flexible, and generally a much more intelligent bus. The ADB bus could support... I think 16 peripherals on the bus, complete with a device ID resolution mechanism, hot swappability, etc. Basically, it did most of what USB did, but a decade earlier, albeit at lower bandwidth and with a decidedly more linear signal path. ;-)
Very nice little bus for its time.... Way, way ahead of its time.... As usual, Apple creates and/or chooses a superior technology and everybody else chooses a less flexible, more CPU-intensive bus because it's cheaper. *Sigh*.
Basically, doing full MPEG decode in the GPU made sense before we had vector units in the CPU, but these days, it isn't enough of an advantage to care. And besides, you usually want to have some software control over the presentation---the ability to float windows over it, etc. That sort of thing doesn't work too well if the GPU is doing all of the decoding for you.
And, of course, HD video is a different animal---different resolution, higher data rate, etc. Even if you have custom hardware to decode SD video and send it to a TV on-the-fly, doing the same for HD is sufficiently harder that it probably isn't worth it, particularly since you have a fast CPU handy already.
PDF? Proprietary? Only if you mean Adobe's implementation. There are thousands of tools out there for generating and viewing PDF content in the open source world. Calling PDF proprietary simply because Adobe doesn't provide a viewer for all platforms would be like calling multicast DNS proprietary because at least initially, stock versions of Rendezvous wouldn't compile under Linux.
Based on that same definition, Postscript is proprietary. Oddly enough, Ghostscript is sometimes known to open encapsulated postscript files generated by Adobe Illustrator that Adobe's own Photoshop can't. When the open source software exceeds the quality and reliability of the reference implementation, it can no longer reasonably be described as proprietary, even if the reference implementation happens to be, IMHO.
That said, I would no more recommend people posting PDF or OOo docs than Word docs, for exactly the same reason. You have to download special software to view it. Even if Firefox had a plug-in in the shipping version, most people wouldn't have that version. For that matter, most people don't use Firefox.
The web is a powerful platform for deployment of information precisely because there are a very limited number of standard formats for contents, and a single standard environment for viewing them. It pisses me off to no end when I see a PDF file without an HTML version alongside it. The last thing I want to do is deal with a whole different environment to view content---whether it's Acrobat or a viewer plug-in makes no difference. Ditto for Word, OOo, etc. (As I always say, "Repeat after me: 'HTML is for Viewing, PDF is for Printing'.")
And I hope I -never- have to read something that some clueless peson uploaded in Postscript again. Yes, there's software for every platform, but no, most people don't have it installed, and it's a pain in the ass to distill to PDF just to view something that's usually mostly plain text anyway. And before you ask, yes, sometimes I have been known to just read the Postscript file in vi.
Bottom line, if in doubt, HTML. If HTML won't work because the person posting it is too anal about formatting... HTML anyway, and post a nice, neat, formatted PDF for the three other people in the world who are as anal as they are. ;-)
</rant>
We now return you to your regularly scheduled discussion of open formats.
That said, it's probably a fool's bet if you actually tried to use that defense. :-)
I would strongly encourage everyone reading this to watch CNN's website and if the story reappears and is as slanted as the one I read, write to them and complain.
Thus, by stopping cancer, you're actually -helping- evolution by ensuring that only the useful and non-fatal mutations continue into the next generation.
I know, I know. IHBT. IHL. HAND.
1. Superfluous; going beyond that which is necessary.
2. Repetitious.
Of course, you can agree or disagree, but it can legitimately be said that your continuation of this thread is superfluous.
(This message moderated score 0: Redunandtly redundant.)
Just my $0.02.