Good grief, if they applied that to regular hard drives, you'd be paying $160 for the drive and over $1000 in music taxes for a 250GB drive! Drives are up to 500GB now, and are expected to be up to a TB in 2006, that would be a $4000 tax!
While they're at it, why don't they just tack on a 10 cent tax per sheet of blank paper...maybe the book industry should claim that the reason sales of books are down is because of Internet file sharing.
That looks like an excellent monitor, with somewhat better specs than the Apple monitor (brightness and contrast) (the Apple 20" Cinema Display is basically the same monitor that is built in to the 20" iMac G5 - costs $799).
Why would I need to survey "all of these people"? It is certainly "one of the things people are paying for", and it is certainly "one of the things Apple is charging extra for". That's like saying that I have to survey every person who buys a laptop to see if they are buying it because it is portable before I can list it being portable as one of the things that people pay extra for it being a laptop.
Why had "Apple better deliver a better video card"? It isn't needed for most people! If you think you need more video capability, then you're right, you shouldn't buy an iMac.
If you don't want Mac OSX, then again, you shouldn't buy an iMac. It isn't worth it to you. For people who do buy the iMac or other Apple computers, that is one of the things that make the price worth it. Obviously, or they wouldn't be buying it. Very few people would buy a Mac just to run Linux or BSD on it. Linus would, but he's not quite your typical computer user.
The 20" iMac screen is widescreen at 100dpi, and it is absolutely gorgeous. At least do a comparison with a widescreen LCD monitor. That "weak 64MB video card" is perfectly adequate for most people - contrary to popular opinion on slashdot, most people use their computers for things other than high-powered gaming.
I'd really like to see the system you can build for so cheap that takes up zero space beyond the display. That's one of the things people are paying for. They're also paying for the ability to run Mac OSX. Yes, 256MB of memory is a bit small (but my Mom hasn't really noticed it). It's also cheap and easy to fix (double it for $40, quadruple it for $140 - I'm going to use the original 256MB from my Dad's machine to double my Mom's).
In OSX, just double-click a.ps file to open it in Preview - you get a short "Converting to PDF", then it displays. You can then save it in PDF format if you like.
There are occasionally PDF documents that require Reader 6.0 or better to be able to see anything, but most things still produce files that don't need anything beyond 4. It's amazing how bloated Reader 6 and 7 are.
Although TextEdit on Mac OSX does handle most Word files, a PDF is so much better for passing something around to be read. Since you can print any file to a PDF (Print Preview just prints to a PDF, then displays it), it is a natural standard for anyone to use on OSX, since it is readable on most platforms.
What is owned with a copyright or patent is not the work itself (the words, the song, the methods, the ideas), but a legal concept that gives you rights. You can sell the copyright to someone, but that doesn't affect the millions of copies of the work you sold to other people. You can sell the patent to someone else, but that doesn't affect the devices using the patented methods, or materials produces using the patented processes. Physical property is completely different. You have absolute rights over that piece of physical property, and if you sell it, you have no rights to what someone else does with it. Physical property doesn't evaporate after a period of time (unless the laws of physics dictate so). If you sell it, you are only selling rights. Is your right to free speech "property"?
The ONLY thing that physical property and "intellectual rights" have in common is that they have economic value because they can be bought and sold.
Give an initial automatic copyright which lasts 50 years from creation or 10 years from first publication (whichever is earlier). Something published between 40 and 50 years after creation would be given the full 10 years. Renewal is required every 10 years after that, for a fee of (n * ((1.5^t) - (1.5^(t-10))) (t in years since first publication) where n is set so that the first renewal at 10 years is an inflation-adjusted 100 dollars (initially n would be about 1.73). No maximum term, except that the amounts after about 50 years start to become pretty high even for something like Mickey Mouse, and by 75 years are ridiculously high, and by 100 years are impossibly high.
Another proposal: for an adapted work to be copyrighted with a new creation date, the original it was based on MUST be made available essentially on a GNU-like "for no more than the actual cost of distribution" to anyone who has an authorized copy of the adapted work (this part will need some work clarifying exactly what is meant, e.g. what form would the "original" of a film negative need to be distributed in to meet this requirement). This would include adaptations/restorations of public domain works, so that others can also adapt or restore them without being encumbered by the "new" copyright. This prevents minor updates to a work to avoid the escalating fee, or to re-establish copyright on something that you failed to renew. You can get copyright on the NEW portion of the work, but you can't get a de facto copyright on something merely because you have the only existing copy of an old work that's entered the public domain.
I'd allow a fairly lenient 15 years additional grace period before the copyright finally expires and it enters the public domain (which means a work unpublished for 50 years could be protected for another 25 years after publication and not have paid any fees). The precise meaning of "published" would also have to be clarified. "Published" SHOULD mean that the "public" has access to it, for a reasonable price.
A slight modification would also allow a mandatory-licensing scheme that was fair - base the renewal fee on the price set for being able to buy a copy for the next 10 years (so set *n* in the above formula to "price per copy").
From what you've been saying, you appear to be confused. I agree that it isn't "viral" in the sense that something is "automatically GPL" as soon as a bit of GPL code touches it, and it isn't viral in the sense that code you write independently can be re-released under terms other than the GPL (so having been once in contact with GPL code doesn't mean it is "infected" - all you have to do is remove the GPL code and you can license it in some other way).
However, you do seem to be missing the point of what a "derivative" is. Including a portion of code copied from a GPL program into yours does indeed require you to release the binaries produced from that source code under the GPL, INCLUDING your own source code. Not just the portion that started out GPL. Not even just the portion of GPL code that you've modified. The ENTIRE PROGRAM. This is based on copyright law, not the GPL. If it is a derivative work, then the entire work is required to be released under the GPL, or not distributed at all. "Mere aggregation" is the term the GPL uses to distinguish a different case, that of a collection of different works. Although copyright law would say that a CD-ROM is a derivative work of each of the pieces (as well as being a work of its own), the GPL specifically allows you to do that without making each SEPARATE work be licensed under the GPL.
The FSF is quite clear on all of this in their FAQs, and if you think the FSF lawyers are saying something different, you're probably misinterpreting them.
Well, what's worse is when you have the badguy.com nameserver respond:
badguy.com IN NS a.gtld-servers.net.
and then in the additional records, you inform them that:
a.gtld-servers.net. IN A 1.2.3.4
Now, the additional A record is not exactly unexpected, it's letting you know where that nameserver can be found - otherwise, you might have to keep chasing it around through multiple levels of queries, or even get into a loop which can't be resolved. However, the real problem is that a.gtld-servers.net is one of the DNS for.com and others, so the next time you do a lookup on some other.com, if you cached that address, you'll look it up from 1.2.3.4 (i.e. the bad guy's server), which of course will lie to you about where to find whatever you're looking up.
The correct thing to do is not cache the "glue" records at all - only cache records from the authoritative nameserver, starting with the root nameservers. If they send you addresses for the nameservers mentioned, don't cache them. To be extra paranoid, don't even follow them unless they are sub-domains of the domain the server is authoritative for, but that can end up in extreme or infinite levels of recursion. Then there's lame delegation...
I haven't even seen a Sacajawea coin since the first week they were released and a couple stores were handing them out as change. I still have that one, somewhere, but I dropped it and haven't seen it for a while.
I haven't bothered with Cylinder/Sector/Head since working with floppy disk file systems in the late '70s, early '80s. With SCSI you don't deal with physical layout at all. The old BIOS/IDE limitation, and the brain-dead PC partitioning scheme, has and continues to needlessly complicate what should be a fairly simple task of bootstrapping a computer. That modern drives still need to pretend to have a physical geometry that has nothing do with the actual layout is ridiculous. Even worse, don't SATA controllers have to pretend to be IDE controllers so old systems can use them without needing new drivers (although that hasn't seemed to work very well), thus carrying the stupidity even further into the future? I love how SATA is ATA over serial lines, where ATA is basically SCSI over IDE, so really SATA is just SCSI over serial lines, which is pretty much what FireWire is. The one advantage for SATA was supposed to be IDE compatibility so new drivers wouldn't be required, yet with that being a failure, it seems that the IDE compatibility might be a liability, unnecessarily complicating the controller card.
It's clear from the 1963 reference that the term was already in use. I see the term as being rather neutral - saying "a bunch of programmers broke into the phone network" doesn't indicate that "programmers" is a negative term referring to people who are prone to breaking into things, and substituting "hacker" in shouldn't change that. That it is seen as a negative term is stupid, but not particularly surprising.
No. If something is currently v2-or-later, you can release a version that is v2-only (by adding code that is v2-only), v3-only (by adding code that is v3-only - the v2-or-later code is compatible), v3-or-later, or keep it all v2-or-later. The original code that was v2-or-later doesn't change, but the modified program you're distributing can be any of those. If you don't want that to happen, then you should release your code as v2-only from the start (which doesn't stop anyone from adding v2-or-later code, but does prevent any v3 code from being combined with your code).
GPL v2 and GPL v3 have to be incompatible, one way or the other or both ways. If they weren't, they'd be the same license. If v3 removes restrictions, then v2-only code is incompatible with it (since the essence of any version of the GPL is that additional restrictions aren't allowed - if it were to include an exception for v2-only code, for example, then it is no better than if you just continue to license under v2-or-later). If v3 adds restrictions, then it is incompatible with v2-only code. If it adds some restrictions and removes others, then they are mutually incompatible. Only v2-or-later code is compatible with v3-only or v3-or-later. Two licenses that both require there be "no other restrictions" must be mutually incompatible unless they have exactly the same restrictions. Only dual-licensing of at least one of them (which "or-later" is) can make them compatible.
However, let's say that GPLv3 has additional restrictions AND additional freedoms - you can't get the additional freedoms unless you agree to the additional restrictions - you either get v2 or v3, not a mixture. In addition, let's say that v3 has only additional restrictions. You're right that it won't affect a v2-or-later-licensed program, UNTIL someone adds in v3-or-later code. Then the project as a whole can only be distributed under the v3 terms. You could strip out the v2 code and release that under v2 terms, or continue using only the original version, but if you want those particular changes you'd have to accept v3 terms.
HD-DVD and Blu-Ray players will probably be in the $500 range at that point, not the $200 range. So will the PS3 itself, probably.
Re:PS3 = End of 2006. HD-DVD players in 3 months!
on
Blu-Ray vs. HD-DVD
·
· Score: 1
The first players are going to be too expensive to be snatched up, particularly with the threat of a format war. HD-DVD may hurt its chances by coming out so early, as it will probably sell so few that even if it is in line with predictions, people will associate it with being a failure. Then when PS3 comes out, and Sony puts out 6000 titles on Blu-Ray discs, the associated publicity with how big of a hit it is will probably tip the balance then and there.
I'm not likely to buy hybrid discs unless they don't cost me any extra. Titles that come out in HD-DVD will be, for me, as if they didn't come out. The industry isn't going to be able to abandon the vast majority of the market that won't be upgrading to HDTV any time soon, much less an expensive player that has a good possibility of being obsolete in a year. The industry probably also won't be able to abandon their hopes and dreams of gouging their customers by charging too much for the new discs.
Blu-Ray hybrid discs are also possible, so that isn't going to be a compelling argument for HD-DVD. If Apple comes out with a Blu-Ray reader, or better yet a writer (that can also do CD-R/RW, DVD+-R/RW) at about the same time that Sony comes out with PS3, I think the game will be over then. Everyone else will jump on the bandwagon and HD-DVD will vanish within a year (e.g. by the end of 2007, if that long).
Linux "comes" in whatever directory you put it in, it could be in/var/log/kernel-source for all you know. You're right that putting in the trailing slash is a more general way of making sure it works even if it is a symbolic link. It has the unfortunate side effect of making the resulting list a bit messier by inserting an extra / in the resulting paths (e.g./usr/src/linux//README) using the BSD version of find, where the proper solution is to use the -H flag. Of course, the more annoying aspect of BSD find is that it requires at least one path, it doesn't default to the current directory. On the other hand, the annoyance with GNU find is that when you omit the directory, it still puts in the./ as part of the listed path. What I really want is a -D option to specify a directory to change to first (best not to use -d, as that is a synonym for -depth in BSD find).
Actually, you're right that mine doesn't work, because I don't even have a symbolic link at/usr/src/linux, much less a directory. However, my example is still be many times faster, if there is a directory full of Linux source at/usr/src/linux - for example, on a dual 1.8GHz G5, with all files cached in memory, against a particular version of Linux with a bit over 4 million lines of.c and.h files consisting of 120 million bytes, running my version took 3.2 seconds real time, 2.7 seconds CPU time, 1.6 seconds system time. Using -exec in find took about 46 seconds real time, 9.5 seconds CPU, 39.1 system. That's running under Mac OSX with the BSD version of find packaged with the system.
To be even more general, you should run find with the -print0 argument, and run xargs with -0, just in case you have a file with embedded naughty characters, and you should also probably put in a check for -type f, in case someone created a directory called files.c or something. Doesn't change the point of my post.
The UK is about the size of Oregon, but Oregon has a very low population density. The UK is similar in shape to Illinois, but is a bit larger. It has around 4-5 times the population of Illinois, about 2.9 times the population density. New Jersey, Rhode Island, Massachusetts and Connecticut all have population densities higher than the UK. There are 11 states larger than the UK (both by total and by land area). California is about 1.7 times the area of the UK, but the UK has a population about 1.7 times that of California, with the UK having a population density about 3 times larger.
UK population is closer to 1/5th than 1/4th the US population. The UK population density is about 7.8 times that of the United States as a whole.
Even more efficient, when "running backwards", go back to the checkpoint and re-execute everything while logging each instruction (log an instruction: record the previous value of anything this instruction overwrites, including the program counter if not the next instruction). Then you can do a direct "go backwards" mode. You still need to log each memory location that gets overwritten between snapshots (and use those values when re-executing).
Instead of snapshots, you could record state at the beginning of each basic block (any location in the code that can be reached from a location not immediately preceding it, i.e. after any branch instruction or interrupt/exception).
For dealing with devices, you could log all interactions with I/O ports (record both read and write values), and either log all DMA activity by the device or log all memory reads from memory used by the I/O device. For multi-processor, you have to emulate all the processors in a deterministic way (preferably as close to the original timing as possible, but since programs shouldn't be sensitive to timing that way, it would be helpful to be able to vary the timing as well - as long as it can be repeated during the re-execution).
There still might be bugs that can't be found this way - things that are highly timing dependent, either with the exact timing between multiple CPUs or with I/O devices.
You have an odd definition of "better" and "worse". Much of what you talk about is "different". Analog is "worse" than the original medium (the original medium being the singer's voice, the drum, the guitar, the flute), where "worse" means "less perfectly reproduced". Digital isn't inherently better or worse than analog at reproduction, it is just "different".
How is it that you define a "physical" musical instrument as better than a synthesized one? Sure, a cheap synthesizer probably sounds worse than most instruments to most people, but if someone learned to listen to music on a good synthesizer, they'd think those old-fashioned wood and metal and string instruments are the "poor substitute for the real thing".
I've seen touched-up scanned photos re-printed on a decent inkjet photo printer that were much better than the original "analog" photo it was scanned from.
PDFs compared to chip Data Books, especially when you can search the PDF? PDF, hands down.
A movie in the theater isn't better than a DVD because the DVD is digital, it's because the theater screen is huge. With enough bandwidth, a digital theater projection would be just as good (well, OK, we still have a ways to go in color saturation and range, but you'd have the same problem with analog - the problem isn't analog vs. digital, but electronic vs. film).
And, while a good travel agent can sometimes find you a good deal or a better trip, you can often do a better job on your own. It's nice to have the choice, and it's good to do a little leg work before you go to the travel agent in the first place, as a check on what they're doing.
So I disagree with your premise that "digitization turns everything into crap". What it does do is turn crap into crap at much faster rates and for cheaper than ever before, so we get lots more of it drowning out the good stuff.
I wouldn't call 22MHz "far higher frequency than human hearing can detect". Yes, it is beyond the range most people can hear, and even if you can hear it, most people can't tell the difference when it is presented with the rest of the frequencies usually present in music, but it isn't "far higher".
I'd really like to see any audiophile presented with a test: play their choice of vinyl record over their choice of turntable, pre-amp, equalizer, amp, monster cables, filtered power supplies, speakers - and let an expert record the output at the pre-amp and put it on CD. Then let the audiophile listen again and figure out which is which (with the CD being fed in at the same pre-amp into the same system). I'll bet some could - and they would rightly be called freaks of nature, and totally irrelevant to the world of consumer music.
$100-150 speakers sound just as good as $500 speakers unless you have a pretty good room to set everything up in. Most people set up some speakers in the bedroom and/or the living room, hook it to a $200 amp and a $100 CD or DVD player - if they have 5.1 surround, they're lucky if they have a place to mount the rear speakers that's anywhere near to where they should be (and plenty of them bought the "5.1 SURROUND SOUND COMPLETE HOME THEATER SYSTEM" for $199.99, so why does it matter?). Why would they blow $1000 on a pair of speakers with a setup like that? Besides, by the time most people can afford a house with a listening room and the gear to go in it, they've probably lost enough of their hearing where they can't tell the difference anyway.
Slashdot Mirror
Good grief, if they applied that to regular hard drives, you'd be paying $160 for the drive and over $1000 in music taxes for a 250GB drive! Drives are up to 500GB now, and are expected to be up to a TB in 2006, that would be a $4000 tax!
While they're at it, why don't they just tack on a 10 cent tax per sheet of blank paper...maybe the book industry should claim that the reason sales of books are down is because of Internet file sharing.
That looks like an excellent monitor, with somewhat better specs than the Apple monitor (brightness and contrast) (the Apple 20" Cinema Display is basically the same monitor that is built in to the 20" iMac G5 - costs $799).
Why would I need to survey "all of these people"? It is certainly "one of the things people are paying for", and it is certainly "one of the things Apple is charging extra for". That's like saying that I have to survey every person who buys a laptop to see if they are buying it because it is portable before I can list it being portable as one of the things that people pay extra for it being a laptop.
Why had "Apple better deliver a better video card"? It isn't needed for most people! If you think you need more video capability, then you're right, you shouldn't buy an iMac.
If you don't want Mac OSX, then again, you shouldn't buy an iMac. It isn't worth it to you. For people who do buy the iMac or other Apple computers, that is one of the things that make the price worth it. Obviously, or they wouldn't be buying it. Very few people would buy a Mac just to run Linux or BSD on it. Linus would, but he's not quite your typical computer user.
The 20" iMac screen is widescreen at 100dpi, and it is absolutely gorgeous. At least do a comparison with a widescreen LCD monitor. That "weak 64MB video card" is perfectly adequate for most people - contrary to popular opinion on slashdot, most people use their computers for things other than high-powered gaming.
I'd really like to see the system you can build for so cheap that takes up zero space beyond the display. That's one of the things people are paying for. They're also paying for the ability to run Mac OSX. Yes, 256MB of memory is a bit small (but my Mom hasn't really noticed it). It's also cheap and easy to fix (double it for $40, quadruple it for $140 - I'm going to use the original 256MB from my Dad's machine to double my Mom's).
In OSX, just double-click a .ps file to open it in Preview - you get a short "Converting to PDF", then it displays. You can then save it in PDF format if you like.
There are occasionally PDF documents that require Reader 6.0 or better to be able to see anything, but most things still produce files that don't need anything beyond 4. It's amazing how bloated Reader 6 and 7 are.
Although TextEdit on Mac OSX does handle most Word files, a PDF is so much better for passing something around to be read. Since you can print any file to a PDF (Print Preview just prints to a PDF, then displays it), it is a natural standard for anyone to use on OSX, since it is readable on most platforms.
What is owned with a copyright or patent is not the work itself (the words, the song, the methods, the ideas), but a legal concept that gives you rights. You can sell the copyright to someone, but that doesn't affect the millions of copies of the work you sold to other people. You can sell the patent to someone else, but that doesn't affect the devices using the patented methods, or materials produces using the patented processes. Physical property is completely different. You have absolute rights over that piece of physical property, and if you sell it, you have no rights to what someone else does with it. Physical property doesn't evaporate after a period of time (unless the laws of physics dictate so). If you sell it, you are only selling rights. Is your right to free speech "property"?
The ONLY thing that physical property and "intellectual rights" have in common is that they have economic value because they can be bought and sold.
Where do you get the idea that glass is anti-static? Glass is an effective insulator and very easy to build up static electricity on it.
Give an initial automatic copyright which lasts 50 years from creation or 10 years from first publication (whichever is earlier). Something published between 40 and 50 years after creation would be given the full 10 years. Renewal is required every 10 years after that, for a fee of (n * ((1.5^t) - (1.5^(t-10))) (t in years since first publication) where n is set so that the first renewal at 10 years is an inflation-adjusted 100 dollars (initially n would be about 1.73). No maximum term, except that the amounts after about 50 years start to become pretty high even for something like Mickey Mouse, and by 75 years are ridiculously high, and by 100 years are impossibly high.
Another proposal: for an adapted work to be copyrighted with a new creation date, the original it was based on MUST be made available essentially on a GNU-like "for no more than the actual cost of distribution" to anyone who has an authorized copy of the adapted work (this part will need some work clarifying exactly what is meant, e.g. what form would the "original" of a film negative need to be distributed in to meet this requirement). This would include adaptations/restorations of public domain works, so that others can also adapt or restore them without being encumbered by the "new" copyright. This prevents minor updates to a work to avoid the escalating fee, or to re-establish copyright on something that you failed to renew. You can get copyright on the NEW portion of the work, but you can't get a de facto copyright on something merely because you have the only existing copy of an old work that's entered the public domain.
I'd allow a fairly lenient 15 years additional grace period before the copyright finally expires and it enters the public domain (which means a work unpublished for 50 years could be protected for another 25 years after publication and not have paid any fees). The precise meaning of "published" would also have to be clarified. "Published" SHOULD mean that the "public" has access to it, for a reasonable price.
A slight modification would also allow a mandatory-licensing scheme that was fair - base the renewal fee on the price set for being able to buy a copy for the next 10 years (so set *n* in the above formula to "price per copy").
From what you've been saying, you appear to be confused. I agree that it isn't "viral" in the sense that something is "automatically GPL" as soon as a bit of GPL code touches it, and it isn't viral in the sense that code you write independently can be re-released under terms other than the GPL (so having been once in contact with GPL code doesn't mean it is "infected" - all you have to do is remove the GPL code and you can license it in some other way).
However, you do seem to be missing the point of what a "derivative" is. Including a portion of code copied from a GPL program into yours does indeed require you to release the binaries produced from that source code under the GPL, INCLUDING your own source code. Not just the portion that started out GPL. Not even just the portion of GPL code that you've modified. The ENTIRE PROGRAM. This is based on copyright law, not the GPL. If it is a derivative work, then the entire work is required to be released under the GPL, or not distributed at all. "Mere aggregation" is the term the GPL uses to distinguish a different case, that of a collection of different works. Although copyright law would say that a CD-ROM is a derivative work of each of the pieces (as well as being a work of its own), the GPL specifically allows you to do that without making each SEPARATE work be licensed under the GPL.
The FSF is quite clear on all of this in their FAQs, and if you think the FSF lawyers are saying something different, you're probably misinterpreting them.
Well, what's worse is when you have the badguy.com nameserver respond:
and then in the additional records, you inform them that:Now, the additional A record is not exactly unexpected, it's letting you know where that nameserver can be found - otherwise, you might have to keep chasing it around through multiple levels of queries, or even get into a loop which can't be resolved. However, the real problem is that a.gtld-servers.net is one of the DNS forThe correct thing to do is not cache the "glue" records at all - only cache records from the authoritative nameserver, starting with the root nameservers. If they send you addresses for the nameservers mentioned, don't cache them. To be extra paranoid, don't even follow them unless they are sub-domains of the domain the server is authoritative for, but that can end up in extreme or infinite levels of recursion. Then there's lame delegation...
I haven't even seen a Sacajawea coin since the first week they were released and a couple stores were handing them out as change. I still have that one, somewhere, but I dropped it and haven't seen it for a while.
I haven't bothered with Cylinder/Sector/Head since working with floppy disk file systems in the late '70s, early '80s. With SCSI you don't deal with physical layout at all. The old BIOS/IDE limitation, and the brain-dead PC partitioning scheme, has and continues to needlessly complicate what should be a fairly simple task of bootstrapping a computer. That modern drives still need to pretend to have a physical geometry that has nothing do with the actual layout is ridiculous. Even worse, don't SATA controllers have to pretend to be IDE controllers so old systems can use them without needing new drivers (although that hasn't seemed to work very well), thus carrying the stupidity even further into the future? I love how SATA is ATA over serial lines, where ATA is basically SCSI over IDE, so really SATA is just SCSI over serial lines, which is pretty much what FireWire is. The one advantage for SATA was supposed to be IDE compatibility so new drivers wouldn't be required, yet with that being a failure, it seems that the IDE compatibility might be a liability, unnecessarily complicating the controller card.
It's clear from the 1963 reference that the term was already in use. I see the term as being rather neutral - saying "a bunch of programmers broke into the phone network" doesn't indicate that "programmers" is a negative term referring to people who are prone to breaking into things, and substituting "hacker" in shouldn't change that. That it is seen as a negative term is stupid, but not particularly surprising.
No. If something is currently v2-or-later, you can release a version that is v2-only (by adding code that is v2-only), v3-only (by adding code that is v3-only - the v2-or-later code is compatible), v3-or-later, or keep it all v2-or-later. The original code that was v2-or-later doesn't change, but the modified program you're distributing can be any of those. If you don't want that to happen, then you should release your code as v2-only from the start (which doesn't stop anyone from adding v2-or-later code, but does prevent any v3 code from being combined with your code).
GPL v2 and GPL v3 have to be incompatible, one way or the other or both ways. If they weren't, they'd be the same license. If v3 removes restrictions, then v2-only code is incompatible with it (since the essence of any version of the GPL is that additional restrictions aren't allowed - if it were to include an exception for v2-only code, for example, then it is no better than if you just continue to license under v2-or-later). If v3 adds restrictions, then it is incompatible with v2-only code. If it adds some restrictions and removes others, then they are mutually incompatible. Only v2-or-later code is compatible with v3-only or v3-or-later. Two licenses that both require there be "no other restrictions" must be mutually incompatible unless they have exactly the same restrictions. Only dual-licensing of at least one of them (which "or-later" is) can make them compatible.
However, let's say that GPLv3 has additional restrictions AND additional freedoms - you can't get the additional freedoms unless you agree to the additional restrictions - you either get v2 or v3, not a mixture. In addition, let's say that v3 has only additional restrictions. You're right that it won't affect a v2-or-later-licensed program, UNTIL someone adds in v3-or-later code. Then the project as a whole can only be distributed under the v3 terms. You could strip out the v2 code and release that under v2 terms, or continue using only the original version, but if you want those particular changes you'd have to accept v3 terms.
Maybe they should work with the MIT Fab Lab people (see their overview, or the Wired article by Bruce Sterling).
HD-DVD and Blu-Ray players will probably be in the $500 range at that point, not the $200 range. So will the PS3 itself, probably.
The first players are going to be too expensive to be snatched up, particularly with the threat of a format war. HD-DVD may hurt its chances by coming out so early, as it will probably sell so few that even if it is in line with predictions, people will associate it with being a failure. Then when PS3 comes out, and Sony puts out 6000 titles on Blu-Ray discs, the associated publicity with how big of a hit it is will probably tip the balance then and there.
I'm not likely to buy hybrid discs unless they don't cost me any extra. Titles that come out in HD-DVD will be, for me, as if they didn't come out. The industry isn't going to be able to abandon the vast majority of the market that won't be upgrading to HDTV any time soon, much less an expensive player that has a good possibility of being obsolete in a year. The industry probably also won't be able to abandon their hopes and dreams of gouging their customers by charging too much for the new discs.
Blu-Ray hybrid discs are also possible, so that isn't going to be a compelling argument for HD-DVD. If Apple comes out with a Blu-Ray reader, or better yet a writer (that can also do CD-R/RW, DVD+-R/RW) at about the same time that Sony comes out with PS3, I think the game will be over then. Everyone else will jump on the bandwagon and HD-DVD will vanish within a year (e.g. by the end of 2007, if that long).
Also, the original was wrong, and not only pedantically: it only gets the .h files. Needs a \( and \) around the -name tags:
with my full-blown version being:Linux "comes" in whatever directory you put it in, it could be in /var/log/kernel-source for all you know. You're right that putting in the trailing slash is a more general way of making sure it works even if it is a symbolic link. It has the unfortunate side effect of making the resulting list a bit messier by inserting an extra / in the resulting paths (e.g. /usr/src/linux//README) using the BSD version of find, where the proper solution is to use the -H flag. Of course, the more annoying aspect of BSD find is that it requires at least one path, it doesn't default to the current directory. On the other hand, the annoyance with GNU find is that when you omit the directory, it still puts in the ./ as part of the listed path. What I really want is a -D option to specify a directory to change to first (best not to use -d, as that is a synonym for -depth in BSD find).
Actually, you're right that mine doesn't work, because I don't even have a symbolic link at /usr/src/linux, much less a directory. However, my example is still be many times faster, if there is a directory full of Linux source at /usr/src/linux - for example, on a dual 1.8GHz G5, with all files cached in memory, against a particular version of Linux with a bit over 4 million lines of .c and .h files consisting of 120 million bytes, running my version took 3.2 seconds real time, 2.7 seconds CPU time, 1.6 seconds system time. Using -exec in find took about 46 seconds real time, 9.5 seconds CPU, 39.1 system. That's running under Mac OSX with the BSD version of find packaged with the system.
To be even more general, you should run find with the -print0 argument, and run xargs with -0, just in case you have a file with embedded naughty characters, and you should also probably put in a check for -type f, in case someone created a directory called files.c or something. Doesn't change the point of my post.
The UK is about the size of Oregon, but Oregon has a very low population density. The UK is similar in shape to Illinois, but is a bit larger. It has around 4-5 times the population of Illinois, about 2.9 times the population density. New Jersey, Rhode Island, Massachusetts and Connecticut all have population densities higher than the UK. There are 11 states larger than the UK (both by total and by land area). California is about 1.7 times the area of the UK, but the UK has a population about 1.7 times that of California, with the UK having a population density about 3 times larger.
UK population is closer to 1/5th than 1/4th the US population. The UK population density is about 7.8 times that of the United States as a whole.
Even more efficient, when "running backwards", go back to the checkpoint and re-execute everything while logging each instruction (log an instruction: record the previous value of anything this instruction overwrites, including the program counter if not the next instruction). Then you can do a direct "go backwards" mode. You still need to log each memory location that gets overwritten between snapshots (and use those values when re-executing).
Instead of snapshots, you could record state at the beginning of each basic block (any location in the code that can be reached from a location not immediately preceding it, i.e. after any branch instruction or interrupt/exception).
For dealing with devices, you could log all interactions with I/O ports (record both read and write values), and either log all DMA activity by the device or log all memory reads from memory used by the I/O device. For multi-processor, you have to emulate all the processors in a deterministic way (preferably as close to the original timing as possible, but since programs shouldn't be sensitive to timing that way, it would be helpful to be able to vary the timing as well - as long as it can be repeated during the re-execution).
There still might be bugs that can't be found this way - things that are highly timing dependent, either with the exact timing between multiple CPUs or with I/O devices.
You have an odd definition of "better" and "worse". Much of what you talk about is "different". Analog is "worse" than the original medium (the original medium being the singer's voice, the drum, the guitar, the flute), where "worse" means "less perfectly reproduced". Digital isn't inherently better or worse than analog at reproduction, it is just "different".
How is it that you define a "physical" musical instrument as better than a synthesized one? Sure, a cheap synthesizer probably sounds worse than most instruments to most people, but if someone learned to listen to music on a good synthesizer, they'd think those old-fashioned wood and metal and string instruments are the "poor substitute for the real thing".
I've seen touched-up scanned photos re-printed on a decent inkjet photo printer that were much better than the original "analog" photo it was scanned from.
PDFs compared to chip Data Books, especially when you can search the PDF? PDF, hands down.
A movie in the theater isn't better than a DVD because the DVD is digital, it's because the theater screen is huge. With enough bandwidth, a digital theater projection would be just as good (well, OK, we still have a ways to go in color saturation and range, but you'd have the same problem with analog - the problem isn't analog vs. digital, but electronic vs. film).
And, while a good travel agent can sometimes find you a good deal or a better trip, you can often do a better job on your own. It's nice to have the choice, and it's good to do a little leg work before you go to the travel agent in the first place, as a check on what they're doing.
So I disagree with your premise that "digitization turns everything into crap". What it does do is turn crap into crap at much faster rates and for cheaper than ever before, so we get lots more of it drowning out the good stuff.
I wouldn't call 22MHz "far higher frequency than human hearing can detect". Yes, it is beyond the range most people can hear, and even if you can hear it, most people can't tell the difference when it is presented with the rest of the frequencies usually present in music, but it isn't "far higher".
I'd really like to see any audiophile presented with a test: play their choice of vinyl record over their choice of turntable, pre-amp, equalizer, amp, monster cables, filtered power supplies, speakers - and let an expert record the output at the pre-amp and put it on CD. Then let the audiophile listen again and figure out which is which (with the CD being fed in at the same pre-amp into the same system). I'll bet some could - and they would rightly be called freaks of nature, and totally irrelevant to the world of consumer music.
$100-150 speakers sound just as good as $500 speakers unless you have a pretty good room to set everything up in. Most people set up some speakers in the bedroom and/or the living room, hook it to a $200 amp and a $100 CD or DVD player - if they have 5.1 surround, they're lucky if they have a place to mount the rear speakers that's anywhere near to where they should be (and plenty of them bought the "5.1 SURROUND SOUND COMPLETE HOME THEATER SYSTEM" for $199.99, so why does it matter?). Why would they blow $1000 on a pair of speakers with a setup like that? Besides, by the time most people can afford a house with a listening room and the gear to go in it, they've probably lost enough of their hearing where they can't tell the difference anyway.