Caveat: I am repeating someone else's opinion, I personally do not have a good enough record player to test it.
A few years ago I was working in moderately high end audio. I happened to be in high quality piano reproductions, but I got to read a lot of comments by audiophiles. Something that is immediately apparent is that there are a few people who really know what they are talking about, and a lot of people who are remarkably good at parroting them, without having a clue what they're saying -- reminds me of alternative medicines actually.
Anyway, one physicist challenged an audiophile to justify their support of records instead of compact discs. Their answer (paraphrased) was that their records will have a greater echo at even number harmonics (2x freq, 4x freq, etc) while conversion of a CD to a waveform will cause it to repeat more at odd number harmonics (3x, 5x,) and they consider the even numbered harmonics more aesthetically appealing.
Now, I don't know if this is complete BS. But it caught my eye because, like you, I had worked out the nyquist frequency and concluded that 24 bit 44.1 had enough accuracy for even the most subtle variations.
I do know that the human ear, particularly in some people, is extremely sensitive. There is a popular anecdote about two violinists arguing over A# vs Bb (needs more context to actually make sense, google if interested). Also, the difference between say a 7th and a 13th (getting a bit more into my field here) causes some really interesting harmonic interactions that really affect the music -- try running some low freq (say 300Hz) music through a filter dropping frequencies over say 15kHz and you'll begin seeing what these people are talking about. Carrying on, there is well studies and very obvious interactions between the harmony of the notes and the melody at F_0 (base freq.), but every intrument has a 'voice' which is the percentage of energy given to every frequency. Specifically the piano has a peak on F_4 (its 5th harmonic, i.e. at 32x the base frequency) -- the amount depending on the piano.
Now, tying back in to the origonal subject. The interactions between these harmonics is one thing which gives music its depth ('mistimings' is another). So I am not willing to write off as uninformed ranting the idea that the different harmonics produced by an analogue device are more pleasing than those produced by a digital device.
No, nv's 2d support is rock solid and works out of the box, but it aint fast.
You won't believe how annoying it is to upgrade your kernel and on reboot be dumped into text mode because of the nvidia driver. However, every every time I go back to nv because I'm pissed at nvidia, I get too frustrated by how slow nv is.
So, I'm back to using nvidia, and every time I upgrade my kernel I have to compile three seperate drivers on reboot to have things work again! (nvidia, vmware, and at76c503c.sf.net -- gpl but not in the stock kernel).
Insurance premiums for doctors are as much as (more?) $100k (depending on area, field, etc). Now, for an ordinary doctor that might not be a problem -- you just put your prices up. Insurance trebles? Put your prices up again. Much the same as white box manufactures don't have to worry overmuch about components fluctuating in price since their compeditors will have the same fluctuations. Do you get it? For the average doctor, this insurance won't affect their income at all.
But there are exceptions and they're not good ones. Imagine if you don't want to work fulltime, perhaps you've retiring or have just had a kid. Suddenly $100k goes from being $40 per billable hour to $80 per billable hour, and you can't compete. Conversely, doctors putting in more hours a week can spread the fixed cost thinner, and really rake in the money.
The premiums haven't changed the likelyhood of lawsuits (which is the goal of a higher price in a free market), instead they've made doctors work longer hours and not have families. Dunno about you, but I don't see that as a good solution to lawsuits with stupidly high payouts.
Oh, and don't think this just applies to malpractice insurance. Doctors get hit with all sorts of stupid bills ($1000 for a radiation licence that must be renewed every year at the same cost -- where the licence is just a piece of paper, no tests or checks!?) As above, this is generally just accepted with a shrug and prices passed along.
Apparently unlike your other readers, I quite liked your article "Know your options". One point you didn't mention is that imdemnification is so rare as to be almost nonexistant throughout the software industry. Microsoft does not imdemnify windows users; Apple does not imdemnify Mac users; Sun does not imdemnify solaris users; IBM does not imdemnify AIX users; and HP does not imdeminfy HPUX users! Who does offer imdemnifiation? Well, some of these companies will imdemnify if the client wants it enough (I only know for certain with HP, but would be unsurprised to find they all will). Given this, it seems more than a little odd to expect any imdemnification of Linux users. I think the only reason anybody is getting linux imdemnification now is that the cost _in this case_ is lower than the PR value.
As for the relative merits of the case, I personally have read an awful lot, and am pretty damn confident. Here's the hurdle race I see SCO needing to run in order to demonstrate that I owe them money on an ongoing basis:
a) SCO owns UNIX rather than Novell. b) SCO manages to overturn AT&T's statement on derivative works and c) convince the judge that AIX is a derivative work d) SCO convinces another judge that since AIX is derivative, any code based on AIX (such as RCU in linux) is therefore derivative. e) Somehow the IBM contrabutions cannot be removed -- perhaps they've been copied to too many places, making it is impossible to ensure all traces are eliminated. f) SCO manages to convince a judge that the GPL is meaningless, rendering it functionally equivilant to the public domain. g) Having succeeded at destroying linux and the GPL, most programs runs on other more obscure operating systems (BSD, Hurd, etc.). After the GPL is rewritten to revalidate it, SCO would have to start again against whichever OS people jump ship to.
For the first hurdle, I'm guessing odds are 50/50. Beyond that I see every single hurdle to be huge, I would consider a case based on even one to be a long short. To try and jump all of them, well winning the lottery six tmes in a row sounds more likely... Having said that, even if SCO gets past the first hurdle they can expect to achieve quite a bit of damage to the linux community.
But the purpose of this letter was to point out that imdemnification just doesn't happen, not to discuss SCO's chances of success at repositioning itself as a linux parasite.
An odd code, which seems to span three continents but only include poorer countries. I wonder if they're trying to avoid piracy by keeping all the poorer countries in the same region?
Given the difficuilty of buying region 5 encoded DVDs, you might be better removing the region coding and remastering it unencrypted.
Personally I take a different view. They are not that different. Getting an email address could be constrewed as opt-in too. The key difference is that robots.txt is really simple to do if you would rather opt out. Much easier than opting out to every single spam.
Incidentially, I know somebody who did exactly that with RHAS and libssl didn't work.
But I was comparing the development model, not the quality of the distributions. RedHat releases a 'community' version to people who want something relatively recent. Lets the community fix the bugs and then releases a 'stable' version. Mandrake now does _exactly_ the same.
Acutally, you have three choices. GPL, QPL or commercial. You can pretty much forget about QPL, which gives GPL and commercial. Oh, in answer to your question, no you can't mix QPL with GPL, that's why the GPLed version of QT was created.
As for the 149 vs 1490 etc, I have only ever worked on individual projects so don't consider anything I say authoritative. According to the trolls, you are supposed to licence everybody who actually uses QT. For a monolithic application (no clean frontend/backend split), I think you would be very hard pressed not to have everybody using QT -- even if it is just for moc and connecting the signals.
Since my style of developing is to create a text based backend, this reduces the number of licences required. There is also site licences.
Anybody wanting to avoid costs could legally (but immorally) do so pretty easily. Start the application as GPLed (if you don't give anybody the binaries you don't have to give the source to anybody). Then you can use QT for free. When you're ready to go commercial, relicence your program and pay for one commerical QT licence.
As for avoiding paying QT entirely in a commerical setting, the short answer is no. You would have to go down the route of selling GPLed code. Of course, this is likely to be fine for a large business developing inhouse code.
I like KDE slightly better... but I would not write commercial software for it.
Have you tried writing commercial software for it? You pay your one-off fee and you can do pretty much anything you like with it. Writing for QT is much easier than writing for GTK+. Just spend some of the money you save.
I simply fail to see the problem -- free software developers are happy with the GPL. Commercial software developers are happy paying a licence. Or are you saying the commerical developers are not happy to pay for the tools they're using? Well, that's ok. They can use what is currently an inferior product (GTK+). Maybe if enough choose this route and they help improve GTK+ it will catch up to QT one day.
I first started using the net in '94. My ISP was sans.vuw.ac.nz. Anyway, everyone was given a free internet account with no monthly fees, but charged by the byte for any traffic at US 10c/MB. Hmm, seems rates have dropped:-)
Every subscriber was also given their own IP (IIRC, mine was 202.20.76.83), and their own matching email address and domainname (lakeland.sans.vuw.ac.nz,lakeland@sans.vuw.ac.nz).
Intel P4 and Xeon beat 4 of the 5 you name on SPEC..
They do? The project I'm working on runs faster on a "AMD Athlon(tm) MP 2000+" than it does on a "Intel(R) Pentium(R) 4 CPU 2.80GHz". Just normal code, in a number of different languages.
You can get such lists pretty easily without having to type them in. Just looking up the most frequently used POS for that word gives almost 90% accuracy. Alternatively I wrote a program that automatically predicts the POS for new words.
However, your BNF grammer is likely to come unstuck as soon as you try to parse either casual english or moderately complex english. Either one very quickly leads to adding lots of infrequently used grammar rules, and hence lots of ambiguity in even simple sentences.
The idea of controlled english was to create a useful subset of english that does conform to a BNF grammar (or LL(1), or something, I forget). Writing in it turns out to be quite hard -- very easy to forget you're writing in a programming language. But there is at least one english controlled english machine-assisted translator.
Given a few years, I wouldn't be surprised to see a program like that be the basis of the next big thing in programming languages.
Actually, I saw someone working on something like parsing english as a programming language, try a Google for 'controlled english' sometime. The general idea is that management may not be able to write the specifications, but they can read them and tell you it isn't what they're really after _before_ you code the thing.
Slashdot Mirror
Caveat: I am repeating someone else's opinion, I personally do not have a good enough record player to test it.
A few years ago I was working in moderately high end audio. I happened to be in high quality piano reproductions, but I got to read a lot of comments by audiophiles. Something that is immediately apparent is that there are a few people who really know what they are talking about, and a lot of people who are remarkably good at parroting them, without having a clue what they're saying -- reminds me of alternative medicines actually.
Anyway, one physicist challenged an audiophile to justify their support of records instead of compact discs. Their answer (paraphrased) was that their records will have a greater echo at even number harmonics (2x freq, 4x freq, etc) while conversion of a CD to a waveform will cause it to repeat more at odd number harmonics (3x, 5x,) and they consider the even numbered harmonics more aesthetically appealing.
Now, I don't know if this is complete BS. But it caught my eye because, like you, I had worked out the nyquist frequency and concluded that 24 bit 44.1 had enough accuracy for even the most subtle variations.
I do know that the human ear, particularly in some people, is extremely sensitive. There is a popular anecdote about two violinists arguing over A# vs Bb (needs more context to actually make sense, google if interested). Also, the difference between say a 7th and a 13th (getting a bit more into my field here) causes some really interesting harmonic interactions that really affect the music -- try running some low freq (say 300Hz) music through a filter dropping frequencies over say 15kHz and you'll begin seeing what these people are talking about. Carrying on, there is well studies and very obvious interactions between the harmony of the notes and the melody at F_0 (base freq.), but every intrument has a 'voice' which is the percentage of energy given to every frequency. Specifically the piano has a peak on F_4 (its 5th harmonic, i.e. at 32x the base frequency) -- the amount depending on the piano.
Now, tying back in to the origonal subject. The interactions between these harmonics is one thing which gives music its depth ('mistimings' is another). So I am not willing to write off as uninformed ranting the idea that the different harmonics produced by an analogue device are more pleasing than those produced by a digital device.
Phew, sorry for going on for so long.
No, nv's 2d support is rock solid and works out of the box, but it aint fast.
You won't believe how annoying it is to upgrade your kernel and on reboot be dumped into text mode because of the nvidia driver. However, every every time I go back to nv because I'm pissed at nvidia, I get too frustrated by how slow nv is.
So, I'm back to using nvidia, and every time I upgrade my kernel I have to compile three seperate drivers on reboot to have things work again! (nvidia, vmware, and at76c503c.sf.net -- gpl but not in the stock kernel).
This is already modded as a troll, but I'll bite.
Insurance premiums for doctors are as much as (more?) $100k (depending on area, field, etc). Now, for an ordinary doctor that might not be a problem -- you just put your prices up. Insurance trebles? Put your prices up again. Much the same as white box manufactures don't have to worry overmuch about components fluctuating in price since their compeditors will have the same fluctuations. Do you get it? For the average doctor, this insurance won't affect their income at all.
But there are exceptions and they're not good ones. Imagine if you don't want to work fulltime, perhaps you've retiring or have just had a kid. Suddenly $100k goes from being $40 per billable hour to $80 per billable hour, and you can't compete. Conversely, doctors putting in more hours a week can spread the fixed cost thinner, and really rake in the money.
The premiums haven't changed the likelyhood of lawsuits (which is the goal of a higher price in a free market), instead they've made doctors work longer hours and not have families. Dunno about you, but I don't see that as a good solution to lawsuits with stupidly high payouts.
Oh, and don't think this just applies to malpractice insurance. Doctors get hit with all sorts of stupid bills ($1000 for a radiation licence that must be renewed every year at the same cost -- where the licence is just a piece of paper, no tests or checks!?) As above, this is generally just accepted with a shrug and prices passed along.
And yet, the same task would take me 10 mins in linux, and 3 days in Novell. I think the problem is experience rather than linux being harder.
Here is my letter:
Hiya,
Apparently unlike your other readers, I quite liked your article "Know your
options". One point you didn't mention is that imdemnification is so rare as
to be almost nonexistant throughout the software industry. Microsoft does
not imdemnify windows users; Apple does not imdemnify Mac users; Sun does not
imdemnify solaris users; IBM does not imdemnify AIX users; and HP does not
imdeminfy HPUX users! Who does offer imdemnifiation? Well, some of these
companies will imdemnify if the client wants it enough (I only know for
certain with HP, but would be unsurprised to find they all will). Given
this, it seems more than a little odd to expect any imdemnification of Linux
users. I think the only reason anybody is getting linux imdemnification now
is that the cost _in this case_ is lower than the PR value.
As for the relative merits of the case, I personally have read an awful lot,
and am pretty damn confident. Here's the hurdle race I see SCO needing to
run in order to demonstrate that I owe them money on an ongoing basis:
a) SCO owns UNIX rather than Novell.
b) SCO manages to overturn AT&T's statement on derivative works and
c) convince the judge that AIX is a derivative work
d) SCO convinces another judge that since AIX is derivative, any code based on
AIX (such as RCU in linux) is therefore derivative.
e) Somehow the IBM contrabutions cannot be removed -- perhaps they've been
copied to too many places, making it is impossible to ensure all traces are
eliminated.
f) SCO manages to convince a judge that the GPL is meaningless, rendering it
functionally equivilant to the public domain.
g) Having succeeded at destroying linux and the GPL, most programs runs on
other more obscure operating systems (BSD, Hurd, etc.). After the GPL is
rewritten to revalidate it, SCO would have to start again against whichever
OS people jump ship to.
For the first hurdle, I'm guessing odds are 50/50. Beyond that I see every
single hurdle to be huge, I would consider a case based on even one to be a
long short. To try and jump all of them, well winning the lottery six tmes
in a row sounds more likely... Having said that, even if SCO gets past the
first hurdle they can expect to achieve quite a bit of damage to the linux
community.
But the purpose of this letter was to point out that imdemnification just
doesn't happen, not to discuss SCO's chances of success at repositioning
itself as a linux parasite.
DVD region is 5.
An odd code, which seems to span three continents but only include poorer countries. I wonder if they're trying to avoid piracy by keeping all the poorer countries in the same region?
Given the difficuilty of buying region 5 encoded DVDs, you might be better removing the region coding and remastering it unencrypted.
Interesting that all the problems are interoperability with proprietary software. There haven't been any problems with the people using linux.
to make sure a virus/trojan didnt find its way on to my wifes
Learn how to use the apostrophe key. Else you might get misunderstood.
Personally I take a different view. They are not that different. Getting an email address could be constrewed as opt-in too. The key difference is that robots.txt is really simple to do if you would rather opt out. Much easier than opting out to every single spam.
Where can I get my "User has no money and wouldn't even click an interesting banner if he did so don't waste bandwidth advertising to him" cookie?
Seems you made a typo:
s/advertising/serving/
Incidentially, I know somebody who did exactly that with RHAS and libssl didn't work.
But I was comparing the development model, not the quality of the distributions. RedHat releases a 'community' version to people who want something relatively recent. Lets the community fix the bugs and then releases a 'stable' version. Mandrake now does _exactly_ the same.
Oh yes, fedora... Well, at least they're making the non-community version available, unlike RHAS.
Acutally, you have three choices. GPL, QPL or commercial. You can pretty much forget about QPL, which gives GPL and commercial. Oh, in answer to your question, no you can't mix QPL with GPL, that's why the GPLed version of QT was created.
As for the 149 vs 1490 etc, I have only ever worked on individual projects so don't consider anything I say authoritative. According to the trolls, you are supposed to licence everybody who actually uses QT. For a monolithic application (no clean frontend/backend split), I think you would be very hard pressed not to have everybody using QT -- even if it is just for moc and connecting the signals.
Since my style of developing is to create a text based backend, this reduces the number of licences required. There is also site licences.
Anybody wanting to avoid costs could legally (but immorally) do so pretty easily. Start the application as GPLed (if you don't give anybody the binaries you don't have to give the source to anybody). Then you can use QT for free. When you're ready to go commercial, relicence your program and pay for one commerical QT licence.
As for avoiding paying QT entirely in a commerical setting, the short answer is no. You would have to go down the route of selling GPLed code. Of course, this is likely to be fine for a large business developing inhouse code.
Hope that helps,
Corrin
Have you tried writing commercial software for it? You pay your one-off fee and you can do pretty much anything you like with it. Writing for QT is much easier than writing for GTK+. Just spend some of the money you save.
I simply fail to see the problem -- free software developers are happy with the GPL. Commercial software developers are happy paying a licence. Or are you saying the commerical developers are not happy to pay for the tools they're using? Well, that's ok. They can use what is currently an inferior product (GTK+). Maybe if enough choose this route and they help improve GTK+ it will catch up to QT one day.
In '99 you say? How about this...
:-)
.
I first started using the net in '94. My ISP was sans.vuw.ac.nz. Anyway, everyone was given a free internet account with no monthly fees, but charged by the byte for any traffic at US 10c/MB. Hmm, seems rates have dropped
Every subscriber was also given their own IP (IIRC, mine was 202.20.76.83), and their own matching email address and domainname (lakeland.sans.vuw.ac.nz,lakeland@sans.vuw.ac.nz)
C'mon, we all know that the week we buy the latest gizmo it will be obsolete.
They do? The project I'm working on runs faster on a "AMD Athlon(tm) MP 2000+" than it does on a "Intel(R) Pentium(R) 4 CPU 2.80GHz". Just normal code, in a number of different languages.
Neither. He was just hoping to remind the parent poster something that most people forget:
"We are a federation of individual states, not one big country. We are a democratic constitutional republic, not a democracy".
Oh, definately a). I'd love a monitor with pixels so fine that it looked as sharp as a high-quality print.
Hmmm, maybe. Though if someone phreaks your telephone line then they're liable for any expenses incurred rather than you.
Of course, all sorts of legal things like 'reasonable care' come into play.
But then, the police around here aren't likely to do anything about a burglary, let alone set up a sting on kiddie porn.
You can get such lists pretty easily without having to type them in. Just looking up the most frequently used POS for that word gives almost 90% accuracy. Alternatively I wrote a program that automatically predicts the POS for new words.
However, your BNF grammer is likely to come unstuck as soon as you try to parse either casual english or moderately complex english. Either one very quickly leads to adding lots of infrequently used grammar rules, and hence lots of ambiguity in even simple sentences.
The idea of controlled english was to create a useful subset of english that does conform to a BNF grammar (or LL(1), or something, I forget). Writing in it turns out to be quite hard -- very easy to forget you're writing in a programming language. But there is at least one english controlled english machine-assisted translator.
Given a few years, I wouldn't be surprised to see a program like that be the basis of the next big thing in programming languages.
Indeed. When I get it down I promise to put up a .torrent. Unfortunately I'm only getting 15Kb/s currently (9 hours remaining)
Actually, I saw someone working on something like parsing english as a programming language, try a Google for 'controlled english' sometime. The general idea is that management may not be able to write the specifications, but they can read them and tell you it isn't what they're really after _before_ you code the thing.
I was in the process of downloading this already. Damn you slashdot!
Follow my example and just leave it wide open. All you lose is that your neighbours share your internet connection sometimes. So what?