'Dunno why everybody is excusing Bowie for his choice as if he were too saintly to criticize. Heck, it hurts me too, being a great fan of his old stuff, but no way is he as technologically challanged as folks are making out. He understands the MP3 scene very well --- he was even on the cover of the UK's Internet magazine about a year ago and featured in a leading article about online music.
Far more likely is that he's just one more greedy author who wants to continue shoring up the old power base of institutions that made him so much money in the past.
If the "greats" don't support us, it's time to support instead the newly emerging bands that don't feel obliged to pay back any such favours to the industry.
Your argument hinges on pigeonholing people into one of two disjoint sets: one the set of people who buy Windows software and who are not aware of free software, and the other the set of people who are aware of free software and who do not buy Windows applications.
Reality is much more complex than that. People cross over between the worlds all the time, they may use more than one computer or dual boot, they may experience both worlds by using computers at home and at work or at college, they have friends, colleagues and professors who tell them about the other world, and they read magazines and they surf the web and find out stuff for themselves.
As a result, there is a very real audience of people who inevitably feel the negative perceptions which result from Troll Tech Windows licensing when buying Windows products. For there not to be such an audience would require all Troll customers to be quite dumb.
I am afraid that your worldview is far too simple. Reality is more complex than that.
It's simple: Troll Tech believe they will make more money from selling/supporting Qt for Windows under the traditional software product model than from selling/supporting it under the alternative free software one.
I assume that they've estimated and factored into their costings the less tangible items such as the negative perceptions caused by the licensing.
The trouble is, of course, that it's so easy to estimate the intangibles incorrectly.
The days of the VAX are of course long gone. The thread was about the PC-based *BSDs, and I was trying to ask about the improvements available in the likes of FreeBSD relative to the old BSD4.x that used to run on Vaxen.
As applied to software, daemons and demons are autonomous processes performing a background role, and inferring anything else from a logo/mascot is stupidity in the extreme.
The symbols of Christianity and Islam look like daggers and other hand weapons. Should we infer that their followers are bloodthirsty murderers? Let's not be silly, please.
If FreeBSD documentation is primarily in the form of man pages and hence is indeed kept reasonably up to date then this is a very big selling point for FreeBSD (and the other *BSDs too?), as far as I am concerned.
The gradual dropping of man pages as primary tech info source in favour of the "info" and HTML formats has made rapid and uniform access to concise technical information a very hit or miss affair in Linux, compared to the near 100% man page coverage of some years ago. Having to rely on format converters is nowhere near as good.
If *BSD hasn't suffered that fate, I see it appearing somewhere around here in the near future.
I liked BSD4.* back in the days of the good ol' Vax machines, and I'd quite like to delve back into that scene.
I presume though that the BSD heart in Free/Open/NetBSD has evolved since those days. Does any *BSD follower have a list of major improvements made over the years that they'd like to share with us here?
Pick it on the basis of subject matter, assuming that you already know what interests you. If you're only vaguely sure of the latter, then choose on the basis of an existing multi-student and well-funded research project, so that you don't have to build all your tools from scratch.
On the other hand, you learn a lot when you have to do it all yourself. Er, well, I did, anyway.
Wherever you go though, it'll be fun, and your ultimate success will depend on you and only you. Good luck!
Our NT boxes often crash while idle
on
Kernel 2.2.12
·
· Score: 1
and I've never seen that happen with any other OS. You can't blame it on power glitches either, because everything is UPS'd, and anyway, the Linux and Solaris equipment is powered from the same sources and it stays up just fine.
And BSODs and running out of virtual memory happen almost as frequently. I can't imagine why your NT machines are stable, but it certainly isn't representative of our (very large and multiple) sites -- I can very easily relate to the slagging off that NT gets for lack of stability, through personal experience.
Slipping new kernels in live
on
Kernel 2.2.12
·
· Score: 1
If all kernel structures were linked up into a well-defined tree and all hardware state changes were recorded there then it wouldn't be all that hard to change the kernel while running live, at least in concept. The trouble is, this wouldn't express the way certain changes are interrelated at a strong or even absolute atomic level -- you'd need checkpointing to guarantee that, plus possibly transaction logging if you want to slip the new kernel in between checkpoints.
That's quite true, but in the absence of someone coming up with more appropriate labels, "upmarket/downmarket" does at least hint at the idea and the problem.
Technical people tend to be somewhat more intelligent than the norm, simply because the subject matter requires at least a degree of mental skill. Why then do a rising number seem to relish wallowing in the level of proficiency in communication expressed in Softley's 'Hackers'?
The sad thing is that, disregarding everything else, Abe's responses showed 1000% more literacy skills than 95% of the material we get on Slashdot these days. It's nice to share teccie values with other people, but for that pleasure we sure do have to slum it in a literary and logical downmarket a lot of the time.
Is Slashdot becoming the MTV of techdom? Heck, we used to deride Usenet, but now it's starting to look decidedly upmarket in comparison. When/. coolness is measured by the frequency of occurrence of "sux" and "u own me", it's time to worry.
If those commercial boxes are up-market, what does that make my latest machine: Abit BP6 m/b, dual Celeron 500, 256M memory, dual 80MHz Ultra2 SCSI controllers, dual 100Tx Ethernet, 6 x 9.1 Gb LVD/Ultra2 drives split across the two SCSI buses, 19" rack chassis with dual redundant power supply.
I currently run Celeron 370's, and the population is rising rapidly.
If the folk at Intel decide to demonstrate their forward-looking approach to CPU design by crippling future Celerons, I will have nothing more to do with them. End of story.
If that's the degree of interest they show in catering for the low-end SMP market, they'll get exactly the same amount of interest back.
Using the same logic, when psychologists think/talk/breathe psychology for more than 4 hours a day then this also indicates addictive behaviour. Would they agree that they should be treated for this affliction, likewise?
The only thing that their prognostication indicates is that they don't understand that the Internet is the medium in which we exist. It's no more an addiction for us than it is an addiction for a fish to stay most of the time in water.
As if the above weren't difficult enough for a cracker, note that you can cross over to the more politically-correct world of symmetric and PK cryptographic algorithms by using either of these as transformation functions in the N*(1:M) N-OTP scheme.
Cascading more than one such algorithm together can produce a weak result, but this is not the case when you cascade one with an ideal OTP (it creates total obscuration of any cryptoanalytic weknesses in the applied algorithm). Of course, it would be pointless to do this in an ideal world, cryptographically speaking, since in theory you cannot improve on a OTP. However, N-OPTs are not ideal OTPs, so such algorithms (especially PK ones) can significantly increase the functional parametrization space.
As always though, remember to keep the key bandwidth substantially greater than the data bandwidth, eg. sample any physical noise infrequently. [Don't forget to attach an alert to your noise generator too so that you know when it stops making a racket: trannies, diodes and your parallel port do die occasionally, and that can be cryptographically embarrassing.]
Thermal noise is so easy to generate (one transistor, a couple of resistors and optionally a diode attached to your parallel port is all you need) that achieving good randomness isn't a problem.
OK, so the noise may be pink rather than white, but you can whiten it by a variety of means, and anyway, pinkness doesn't matter as far as OTP algorithms are concerned: you still can't infer the next bit in a sequence as long as your key bandwidth is (substantially) greater than your data bandwidth. Given that proviso, knowing that the noise is pink doesn't help the cracker at all.
The OTP distribution problem is simple to solve if you are willing to compromise on the "oneness" of your pad. Just use a very big highly random pad which you distribute to your partners in freedom (totally secure distribution is not essential), and then apply varying transformation functions to it to generate an unlimited number of Nearly-OTPs.
The transformation functions should be parametrized by N*(1:M) keysets where M is very large: for example, if M is the number of bytes on a publicly-available music or games CD and N is the number of CDs you decide to use, then the distribution problem boils down to sending N names of CDs from Bob to Alice, plus the name or number of a transformation function if you choose to vary that as well.
That's a trivial amount to distribute securely, even over a low-bandwidth covert channel, and the largeness of such an M creates a combinatorial explosion that would give the biggest of cracking computers something to think about for a long time. Consider the number of different CDs in the world (and MP3s!) and the computational task of combining them even for a trivial linear correlation, then consider that N is unknown to the cracker, and then consider that arbitrary transformations can be applied to the keysets first to render linear correlation fruitless: the problem becomes utterly infeasible to solve by brute force, not because of strong cryptographic properties but because the universe of key-seeded transformations on the underlying pad is limited only by the imagination.
So, if you're willing to be pragmatic about it, (N-)OTPs are live and well despite a theoretical key distribution problem --- ie. the ability to dereference from a name to a keyset means that the CD publishers perform the bandwidth-intensive distribution for you.
Darn. After examining the TNT2 reviews on several web sites and discovering that the board from Hercules was widely acclaimed as the best of the lot, a couple of us here already had our wallets open and were waiting for it to come into stock at our local supplier.
What a shame for a good company that's been with the PC market since the earliest days.
Because Linux competes with Solaris, most people seem to think that there is a high chance that Sun will drop the Linux port of StarOffice. But, that's not the only interpretation on the cards.
Linux also promotes the case of Unix versus NT, so it's good for Solaris and hence also for Sun. Following this logic, Sun could well give themselves a shot in the arm at the bottom end of the market by adopting the Linux port for Sparc and selling their workstations and servers with it preloaded, at less cost than with Solaris. This would create a loyal following of Linux/Sparc users, some of whom will promote the case of Sun equipment in their companies, so there is definitely a plus side to Linux in the Sun portfolio.
Sadly, because there are two sides to this coin, there is probably a massive political argument within Sun about it. Sigh.
Just because authors and the music industry made huge profits in the past does not mean that they *have* to continue doing so in the future as a law of nature.
For several decades, replication and distribution of music was hard, something that only a well-funded mega-industry could do, and that process made people a lot of money. Now anyone can do it, for peanuts --- the rules that held before no longer apply, and the natural thing to happen to that money-making process and to the industry that goes with it is for it to die.
The horse carriage industry used to be massive, a backbone of everyday life and a very important source of income for hundreds of thousands, yet now it's dead except as a niche tourist concern. So what? Times change, and just because you've been coining it in for decades doesn't mean that you have the intrinsic right to continue doing so.
I seem to recall that it's only the Mircosoft MTU handling that breaks totally when the packet-too-big responses are blocked: the RFC says that in the absence of a response then the MTU should be decreased to the guaranteed low value for all TCP/IP implementations, but apparently Mircosoft's stack just sits there forever expecting the world to adapt to what it wants.
So, the poor sod running Mircosoft gear can't even get a slow email out asking the offending ISP to allow packet-too-big through... Sigh.
Slashdot Mirror
'Dunno why everybody is excusing Bowie for his choice as if he were too saintly to criticize. Heck, it hurts me too, being a great fan of his old stuff, but no way is he as technologically challanged as folks are making out. He understands the MP3 scene very well --- he was even on the cover of the UK's Internet magazine about a year ago and featured in a leading article about online music.
Far more likely is that he's just one more greedy author who wants to continue shoring up the old power base of institutions that made him so much money in the past.
If the "greats" don't support us, it's time to support instead the newly emerging bands that don't feel obliged to pay back any such favours to the industry.
Somebody asked this question earlier but I don't think there was any response, not even speculation.
Your argument hinges on pigeonholing people into one of two disjoint sets: one the set of people who buy Windows software and who are not aware of free software, and the other the set of people who are aware of free software and who do not buy Windows applications.
Reality is much more complex than that. People cross over between the worlds all the time, they may use more than one computer or dual boot, they may experience both worlds by using computers at home and at work or at college, they have friends, colleagues and professors who tell them about the other world, and they read magazines and they surf the web and find out stuff for themselves.
As a result, there is a very real audience of people who inevitably feel the negative perceptions which result from Troll Tech Windows licensing when buying Windows products. For there not to be such an audience would require all Troll customers to be quite dumb.
I am afraid that your worldview is far too simple. Reality is more complex than that.
It's simple: Troll Tech believe they will make more money from selling/supporting Qt for Windows under the traditional software product model than from selling/supporting it under the alternative free software one.
I assume that they've estimated and factored into their costings the less tangible items such as the negative perceptions caused by the licensing.
The trouble is, of course, that it's so easy to estimate the intangibles incorrectly.
Sorry for the ambiguity of my question.
The days of the VAX are of course long gone. The thread was about the PC-based *BSDs, and I was trying to ask about the improvements available in the likes of FreeBSD relative to the old BSD4.x that used to run on Vaxen.
I hope that's a bit clearer.
Eeek.
As applied to software, daemons and demons are autonomous processes performing a background role, and inferring anything else from a logo/mascot is stupidity in the extreme.
The symbols of Christianity and Islam look like daggers and other hand weapons. Should we infer that their followers are bloodthirsty murderers? Let's not be silly, please.
If FreeBSD documentation is primarily in the form of man pages and hence is indeed kept reasonably up to date then this is a very big selling point for FreeBSD (and the other *BSDs too?), as far as I am concerned.
The gradual dropping of man pages as primary tech info source in favour of the "info" and HTML formats has made rapid and uniform access to concise technical information a very hit or miss affair in Linux, compared to the near 100% man page coverage of some years ago. Having to rely on format converters is nowhere near as good.
If *BSD hasn't suffered that fate, I see it appearing somewhere around here in the near future.
I liked BSD4.* back in the days of the good ol' Vax machines, and I'd quite like to delve back into that scene.
I presume though that the BSD heart in Free/Open/NetBSD has evolved since those days. Does any *BSD follower have a list of major improvements made over the years that they'd like to share with us here?
Pick it on the basis of subject matter, assuming that you already know what interests you. If you're only vaguely sure of the latter, then choose on the basis of an existing multi-student and well-funded research project, so that you don't have to build all your tools from scratch.
On the other hand, you learn a lot when you have to do it all yourself. Er, well, I did, anyway.
Wherever you go though, it'll be fun, and your ultimate success will depend on you and only you. Good luck!
and I've never seen that happen with any other OS. You can't blame it on power glitches either, because everything is UPS'd, and anyway, the Linux and Solaris equipment is powered from the same sources and it stays up just fine.
And BSODs and running out of virtual memory happen almost as frequently. I can't imagine why your NT machines are stable, but it certainly isn't representative of our (very large and multiple) sites -- I can very easily relate to the slagging off that NT gets for lack of stability, through personal experience.
If all kernel structures were linked up into a well-defined tree and all hardware state changes were recorded there then it wouldn't be all that hard to change the kernel while running live, at least in concept. The trouble is, this wouldn't express the way certain changes are interrelated at a strong or even absolute atomic level -- you'd need checkpointing to guarantee that, plus possibly transaction logging if you want to slip the new kernel in between checkpoints.
Hmmm, an interesting problem though.
That's quite true, but in the absence of someone coming up with more appropriate labels, "upmarket/downmarket" does at least hint at the idea and the problem.
Technical people tend to be somewhat more intelligent than the norm, simply because the subject matter requires at least a degree of mental skill. Why then do a rising number seem to relish wallowing in the level of proficiency in communication expressed in Softley's 'Hackers'?
The sad thing is that, disregarding everything else, Abe's responses showed 1000% more literacy skills than 95% of the material we get on Slashdot these days. It's nice to share teccie values with other people, but for that pleasure we sure do have to slum it in a literary and logical downmarket a lot of the time.
/. coolness is measured by the frequency of occurrence of "sux" and "u own me", it's time to worry.
Is Slashdot becoming the MTV of techdom? Heck, we used to deride Usenet, but now it's starting to look decidedly upmarket in comparison. When
If those commercial boxes are up-market, what does that make my latest machine: Abit BP6 m/b, dual Celeron 500, 256M memory, dual 80MHz Ultra2 SCSI controllers, dual 100Tx Ethernet, 6 x 9.1 Gb LVD/Ultra2 drives split across the two SCSI buses, 19" rack chassis with dual redundant power supply.
:-)
I'm in heaven.
The dual Celeron 500's in my BP6 work just fine. It's nice to run a cumulative total of 1GHz. :-)
These socket-370 Celerons are multiplier-locked. I never overclock anything though --- I just don't have time for worrying about instability.
I currently run Celeron 370's, and the population is rising rapidly.
If the folk at Intel decide to demonstrate their forward-looking approach to CPU design by crippling future Celerons, I will have nothing more to do with them. End of story.
If that's the degree of interest they show in catering for the low-end SMP market, they'll get exactly the same amount of interest back.
Using the same logic, when psychologists think/talk/breathe psychology for more than 4 hours a day then this also indicates addictive behaviour. Would they agree that they should be treated for this affliction, likewise?
The only thing that their prognostication indicates is that they don't understand that the Internet is the medium in which we exist. It's no more an addiction for us than it is an addiction for a fish to stay most of the time in water.
As if the above weren't difficult enough for a cracker, note that you can cross over to the more politically-correct world of symmetric and PK cryptographic algorithms by using either of these as transformation functions in the N*(1:M) N-OTP scheme.
Cascading more than one such algorithm together can produce a weak result, but this is not the case when you cascade one with an ideal OTP (it creates total obscuration of any cryptoanalytic weknesses in the applied algorithm). Of course, it would be pointless to do this in an ideal world, cryptographically speaking, since in theory you cannot improve on a OTP. However, N-OPTs are not ideal OTPs, so such algorithms (especially PK ones) can significantly increase the functional parametrization space.
As always though, remember to keep the key bandwidth substantially greater than the data bandwidth, eg. sample any physical noise infrequently. [Don't forget to attach an alert to your noise generator too so that you know when it stops making a racket: trannies, diodes and your parallel port do die occasionally, and that can be cryptographically embarrassing.]
Thermal noise is so easy to generate (one transistor, a couple of resistors and optionally a diode attached to your parallel port is all you need) that achieving good randomness isn't a problem.
OK, so the noise may be pink rather than white, but you can whiten it by a variety of means, and anyway, pinkness doesn't matter as far as OTP algorithms are concerned: you still can't infer the next bit in a sequence as long as your key bandwidth is (substantially) greater than your data bandwidth. Given that proviso, knowing that the noise is pink doesn't help the cracker at all.
The OTP distribution problem is simple to solve if you are willing to compromise on the "oneness" of your pad. Just use a very big highly random pad which you distribute to your partners in freedom (totally secure distribution is not essential), and then apply varying transformation functions to it to generate an unlimited number of Nearly-OTPs.
The transformation functions should be parametrized by N*(1:M) keysets where M is very large: for example, if M is the number of bytes on a publicly-available music or games CD and N is the number of CDs you decide to use, then the distribution problem boils down to sending N names of CDs from Bob to Alice, plus the name or number of a transformation function if you choose to vary that as well.
That's a trivial amount to distribute securely, even over a low-bandwidth covert channel, and the largeness of such an M creates a combinatorial explosion that would give the biggest of cracking computers something to think about for a long time. Consider the number of different CDs in the world (and MP3s!) and the computational task of combining them even for a trivial linear correlation, then consider that N is unknown to the cracker, and then consider that arbitrary transformations can be applied to the keysets first to render linear correlation fruitless: the problem becomes utterly infeasible to solve by brute force, not because of strong cryptographic properties but because the universe of key-seeded transformations on the underlying pad is limited only by the imagination.
So, if you're willing to be pragmatic about it, (N-)OTPs are live and well despite a theoretical key distribution problem --- ie. the ability to dereference from a name to a keyset means that the CD publishers perform the bandwidth-intensive distribution for you.
Darn. After examining the TNT2 reviews on several web sites and discovering that the board from Hercules was widely acclaimed as the best of the lot, a couple of us here already had our wallets open and were waiting for it to come into stock at our local supplier.
What a shame for a good company that's been with the PC market since the earliest days.
But the Mac interface severely limits the power user.
Check out the research article at
http://www.acm.org/cacm/AUG96/antimac.htm .
Because Linux competes with Solaris, most people seem to think that there is a high chance that Sun will drop the Linux port of StarOffice. But, that's not the only interpretation on the cards.
Linux also promotes the case of Unix versus NT, so it's good for Solaris and hence also for Sun. Following this logic, Sun could well give themselves a shot in the arm at the bottom end of the market by adopting the Linux port for Sparc and selling their workstations and servers with it preloaded, at less cost than with Solaris. This would create a loyal following of Linux/Sparc users, some of whom will promote the case of Sun equipment in their companies, so there is definitely a plus side to Linux in the Sun portfolio.
Sadly, because there are two sides to this coin, there is probably a massive political argument within Sun about it. Sigh.
Just because authors and the music industry made huge profits in the past does not mean that they *have* to continue doing so in the future as a law of nature.
For several decades, replication and distribution of music was hard, something that only a well-funded mega-industry could do, and that process made people a lot of money. Now anyone can do it, for peanuts --- the rules that held before no longer apply, and the natural thing to happen to that money-making process and to the industry that goes with it is for it to die.
The horse carriage industry used to be massive, a backbone of everyday life and a very important source of income for hundreds of thousands, yet now it's dead except as a niche tourist concern. So what? Times change, and just because you've been coining it in for decades doesn't mean that you have the intrinsic right to continue doing so.
I seem to recall that it's only the Mircosoft MTU handling that breaks totally when the packet-too-big responses are blocked: the RFC says that in the absence of a response then the MTU should be decreased to the guaranteed low value for all TCP/IP implementations, but apparently Mircosoft's stack just sits there forever expecting the world to adapt to what it wants.
... Sigh.
So, the poor sod running Mircosoft gear can't even get a slow email out asking the offending ISP to allow packet-too-big through