The thing to do isn't to get a patent, but to publish the technology instead, so that it can't be patented at all.
A central repository of prior art would be useful, as somewhere to refer to for help in settling bogus patent claims. www.prior-art.org and www.prior-art.com both exist, but don't seem to contain much at the moment, so they may or may not be relevant.
The distro couldn't be released under the BSD license, since so much of the software in Debian (not just the kernel) is under the GPL. It couldn't be "pinched for a commercial product", except that the FreeBSD kernel and libc could be seperately, like they can now.
Porting Debian to FreeBSD seems like a good way of getting software to be portable to BSDs as well as Linux. Furthermore it would increase the distribution of dpkg and.deb packages, which can't be a bad thing. I'd like.deb to be the "default" package format instead of RPM.:-)
There's already work being done on a Debian/GNU Hurd distribution, which in some ways is more different than FreeBSD (although it does use glibc, which helps).
No, I believe that ssh uses convention crypto for the encrypting the stream.
The public key (RSA) stuff is used for authentication and exchanging the actual keys which are used to encrypt the stream (with DES, blowfish, IDEA, whatever).
The reason I use Perl for CGI script is just that Perl makes text and string handling (which a large proportion of CGI scripts want to do) very easy. Plus it has interfaces to everything else you might need.
C++ makes things a bit easier than C with std::string, but it's still nowhere near Perl for ease of use.
That has the problem that you can't revoke the keys if they're compromised - that's why I'd store the key somewhere you can only get at by proving that you're you somehow. Still not perfect.
If/when using cryptography becomes widespread and everyone's mother uses it, I see the main problem as keeping the secret keys secret, while not losing them/forgetting the passphrase etc. What do you see as viable solutions to this problem?
Possibilities I can think of right now are:
Keys stored on the person (eg jewelry, implants, whatever)
Keys are encrypted/hidden behind some kind of biometric "lock"
People adapt, and find it just as natural to deal with their cryptographic keys as their car and house keys
It's obvious they've put at least some thought into it. Translating to some universal middle-language is clearly the only sensible way to handle translating between many different languages - you only need N two-way translators instead of N^2.
They describe a UNL editor which translates your text into UNL and back again, so that you can check that the translation will be ok as you write it. This sounds like an excellent way to check the results and minimise the errors/inaccuracy.
Actually coming up with a representation which copes with the meaning in all of the other languages is surely a massively difficult task. *If* they manage to solve that, then I'd think that computer (written) natural language recognition is all but solved. (I don't know what the current state-of-the-art is like)
The one thing which annoys me is their insistence on using the terms "enconverter" and "deconverter". I mean the latter sounds almost ok, but "enconverter"? From a group who need to be expert in languages? Yuck!
I wonder what they'll say in the announcement about the fixed (hopefully!) license. Will it be "Oops, we sent out the wrong license, we meant to use this one," will it be "We were unaware of the conflict with the GPL," or "We still think we're right, but here's a new license to appease people"?
Mother nature has prior art for human genes, which will of course be very similar or identical to that of many animals. I don't quite understand how the questions comes up at all.
Now if someone came up with a brand new gene which did something useful, then that should fall under the same roof as software patents, IMO. That argument's been done to death.
The "C" name comment was an irrelevant aside about my own personal confusion. You see when Corel announced their intention to build a Linux distribution, I thought "Wait, what will they do with OpenLinux?" as I got the name confused. It was probably Monday morning at the time, you see.
This could be the first real test of whether the GPL and other software licences are really enforceable. I see three possible outcomes:
1) Non-enforceable. The whole software industry (including free software) goes into disarray. The terms of the GPL would probably be observed by most companies anyway to keep the community goodwill going. Flamewars about the BSD/GPL flame wars will be meaningless, and everyone will go back to coding.
2) Enforceable. Nothing much changes, except that everyone gets a warm fuzzy feeling about the power of their favourite software licence.
3) Corel backs down. This is the only really likely scenario. This all blows over and the only people who still care would use plain Debian instead of Corel's effort anyway.
(I think I've finally got un-confused about Corel and Caldera both being companies starting with 'C' with Linux distributions - Yay!)
For a start, you can't just accelerate to near light speed in an instant. If you want to survive the experience you'll have to keep to an acceleration of a few g. I haven't done the sums (any takers? I'd be quite interested to know the answer) but I wouldn't be surprised if that rather limited the distance you could get to even taking into account the time dilation effects.
Secondly, just think of the amount of energy you'd need to generate for that sort of acceleration (and probably slowing down at the end too, if there's no-one there already to catch you).
...is that the statistical probability of the planets we found has life is way too low, as well as the probability of that life being intelligent, and the probability that It has found how to generate RF signals of a reasonable strength...
...is fairly irrelevant as far as pointing radio telescopes at them. After all, what's the chance of them having build an optical laser powerful enough to be seen on Earth? Pretty negligible. In any case, it's the start they're looking at from this distance rather than the planet itself.
and an even better question, is anyone pointing radio telescopes at these flying rocks?
If you look at the picture at the top of the BBC article, you'll see that it isn't from an optical telescope. I would bet that they _did_ use radio telescopes rather than X-ray, IR, etc., particularly as the observatories are on Earth (Australia and New Zealand) rather than in orbit.
While it's nice to have this sort of evidence of planets around other stars, it's hard to actually really believe it without seeing real pictures of the planet itself.
It's somehow quite depressing that we're extremely unlikely to visit these other solar systems within our lifetime, or even the next few centuries. It's hard to get your head around the idea that even at the fastest speed possible, it may take millennia to get there.
I'd quite like to see the day when (if) we have colonies in other systems, or even other planets in this solar system. It's frustrating seeing it so far in the future.
I should probably stop dreaming and read/watch some more SF.:-)
None of the articles I've seen mention an IR port, which I would find lacking. Having seen how easily you can send info between Psion 5s and PalmIIIs, I wouldn't want to have to go without it.
If the visor doesn't include IR, then I imagine that would be one of the first Springboard gadgets.
I think they really should have come up with a brand new name for the '7 - it's targetted at a different market to the Series 3/Series 5 which came before it. From the pictures of it you can't easily tell that it's much bigger either - it's roughly the same shape.
I wouldn't say no to one if it were offered, of course, but until they come out with jeans with jumbo pockets this one's not for me.
I think you're misunderstanding the term NP-complete.
There are a few relevant problem categories:
P: can be solved in polynomial time. These are "easy".
NP: If you know the answer, you can check that it's right in polynomial time. Working out the answer from scratch may take an exponential amount of time, though. All P problems are also NP. These problems can be solved in polynomial time by a non-deterministic computer.
NP Complete: A subset of NP problems which have an interesting property. If you can find an algorithm to solve any NP complete problem, then you can use that to solve _any other NP problem_ in polynomial time.
So NP Complete problems are the hardest NP problems, but there are other problems which are much harder. Some problems have been proven to be harder than any exponential. That's pretty hard.
Occasionally there's a stack of boxes labelled "BeOS 4.5" or something similar, often with the Dust Puppy sitting on top. What makes you decide what names to put on the boxes? Is it your favourite stuff, advertising/product placement, or what?
Slashdot Mirror
The thing to do isn't to get a patent, but to publish the technology instead, so that it can't be patented at all.
A central repository of prior art would be useful, as somewhere to refer to for help in settling bogus patent claims. www.prior-art.org and www.prior-art.com both exist, but don't seem to contain much at the moment, so they may or may not be relevant.
The distro couldn't be released under the BSD license, since so much of the software in Debian (not just the kernel) is under the GPL. It couldn't be "pinched for a commercial product", except that the FreeBSD kernel and libc could be seperately, like they can now.
Porting Debian to FreeBSD seems like a good way of getting software to be portable to BSDs as well as Linux. Furthermore it would increase the distribution of dpkg and .deb packages, which can't be a bad thing. I'd like .deb to be the "default" package format instead of RPM. :-)
There's already work being done on a Debian/GNU Hurd distribution, which in some ways is more different than FreeBSD (although it does use glibc, which helps).
No, I believe that ssh uses convention crypto for the encrypting the stream.
The public key (RSA) stuff is used for authentication and exchanging the actual keys which are used to encrypt the stream (with DES, blowfish, IDEA, whatever).
The reason I use Perl for CGI script is just that Perl makes text and string handling (which a large proportion of CGI scripts want to do) very easy. Plus it has interfaces to everything else you might need.
C++ makes things a bit easier than C with std::string, but it's still nowhere near Perl for ease of use.
Have you contributed to any free software projects, either at work or in your free time?
Did you use Linux before you started to use it for this job?
That has the problem that you can't revoke the keys if they're compromised - that's why I'd store the key somewhere you can only get at by proving that you're you somehow. Still not perfect.
Possibilities I can think of right now are:
Any thoughts?
It's obvious they've put at least some thought into it. Translating to some universal middle-language is clearly the only sensible way to handle translating between many different languages - you only need N two-way translators instead of N^2.
They describe a UNL editor which translates your text into UNL and back again, so that you can check that the translation will be ok as you write it. This sounds like an excellent way to check the results and minimise the errors/inaccuracy.
Actually coming up with a representation which copes with the meaning in all of the other languages is surely a massively difficult task. *If* they manage to solve that, then I'd think that computer (written) natural language recognition is all but solved. (I don't know what the current state-of-the-art is like)
The one thing which annoys me is their insistence on using the terms "enconverter" and "deconverter". I mean the latter sounds almost ok, but "enconverter"? From a group who need to be expert in languages? Yuck!
Which of the coders working on open source projects do you admire the most? A particular big name like Linus, or someone less well-known?
Should you ever have to step out of the OS picture for whatever reason, who would you (a) chooose, and (b) choose to take your place?
I wonder what they'll say in the announcement about the fixed (hopefully!) license. Will it be "Oops, we sent out the wrong license, we meant to use this one," will it be "We were unaware of the conflict with the GPL," or "We still think we're right, but here's a new license to appease people"?
:-)
Bets anyone?
Mother nature has prior art for human genes, which will of course be very similar or identical to that of many animals. I don't quite understand how the questions comes up at all.
Now if someone came up with a brand new gene which did something useful, then that should fall under the same roof as software patents, IMO. That argument's been done to death.
I didn't say anything about NDAs.
The "C" name comment was an irrelevant aside about my own personal confusion. You see when Corel announced their intention to build a Linux distribution, I thought "Wait, what will they do with OpenLinux?" as I got the name confused. It was probably Monday morning at the time, you see.
This could be the first real test of whether the GPL and other software licences are really enforceable. I see three possible outcomes:
1) Non-enforceable. The whole software industry (including free software) goes into disarray. The terms of the GPL would probably be observed by most companies anyway to keep the community goodwill going. Flamewars about the BSD/GPL flame wars will be meaningless, and everyone will go back to coding.
2) Enforceable. Nothing much changes, except that everyone gets a warm fuzzy feeling about the power of their favourite software licence.
3) Corel backs down. This is the only really likely scenario. This all blows over and the only people who still care would use plain Debian instead of Corel's effort anyway.
(I think I've finally got un-confused about Corel and Caldera both being companies starting with 'C' with Linux distributions - Yay!)
I still think that's rather unlikely.
For a start, you can't just accelerate to near light speed in an instant. If you want to survive the experience you'll have to keep to an acceleration of a few g. I haven't done the sums (any takers? I'd be quite interested to know the answer) but I wouldn't be surprised if that rather limited the distance you could get to even taking into account the time dilation effects.
Secondly, just think of the amount of energy you'd need to generate for that sort of acceleration (and probably slowing down at the end too, if there's no-one there already to catch you).
...is that the statistical probability of the planets we found has life is way too low, as well as the probability of that life being intelligent, and the probability that It has found how to generate RF signals of a reasonable strength...
...is fairly irrelevant as far as pointing radio telescopes at them. After all, what's the chance of them having build an optical laser powerful enough to be seen on Earth? Pretty negligible.
In any case, it's the start they're looking at from this distance rather than the planet itself.
and an even better question, is anyone pointing radio telescopes at these flying rocks?
If you look at the picture at the top of the BBC article, you'll see that it isn't from an optical telescope. I would bet that they _did_ use radio telescopes rather than X-ray, IR, etc., particularly as the observatories are on Earth (Australia and New Zealand) rather than in orbit.
While it's nice to have this sort of evidence of planets around other stars, it's hard to actually really believe it without seeing real pictures of the planet itself.
:-)
It's somehow quite depressing that we're extremely unlikely to visit these other solar systems within our lifetime, or even the next few centuries. It's hard to get your head around the idea that even at the fastest speed possible, it may take millennia to get there.
I'd quite like to see the day when (if) we have colonies in other systems, or even other planets in this solar system. It's frustrating seeing it so far in the future.
I should probably stop dreaming and read/watch some more SF.
None of the articles I've seen mention an IR port, which I would find lacking. Having seen how easily you can send info between Psion 5s and PalmIIIs, I wouldn't want to have to go without it.
If the visor doesn't include IR, then I imagine that would be one of the first Springboard gadgets.
I think they really should have come up with a brand new name for the '7 - it's targetted at a different market to the Series 3/Series 5 which came before it. From the pictures of it you can't easily tell that it's much bigger either - it's roughly the same shape.
I wouldn't say no to one if it were offered, of course, but until they come out with jeans with jumbo pockets this one's not for me.
I think you're misunderstanding the term NP-complete.
There are a few relevant problem categories:
P: can be solved in polynomial time. These are "easy".
NP: If you know the answer, you can check that it's right in polynomial time. Working out the answer from scratch may take an exponential amount of time, though. All P problems are also NP. These problems can be solved in polynomial time by a non-deterministic computer.
NP Complete: A subset of NP problems which have an interesting property. If you can find an algorithm to solve any NP complete problem, then you can use that to solve _any other NP problem_ in polynomial time.
So NP Complete problems are the hardest NP problems, but there are other problems which are much harder. Some problems have been proven to be harder than any exponential. That's pretty hard.
It works for me in Netscape 4.08 and Mozilla M8.
Of course I keep Java and Javascript disabled. Solves more problems than it creates.
Chris
Occasionally there's a stack of boxes labelled "BeOS 4.5" or something similar, often with the Dust Puppy sitting on top. What makes you decide what names to put on the boxes? Is it your favourite stuff, advertising/product placement, or what?
Chris