As far as I understand the issue is a performance one.
For example, scheme is hard to compile towards the JVM or standard C because both lack proper tail call implementation (poping the parameters in the stack before calling a function in tail position).
You can of course circumvent this because as you say the target language is Turing complete, but you lose a lot on performance if the target isn't tailored to this particular use.
I've written a few rudimentary scheme interpreters and have ideas for a new one.
I'm suprised I've never seen scheme and JIT compiling mentionned together.
First, it seems JIT compiling would elegantly solve the problems linked to eval and macros (and their specific environments). When you program a simple scheme interpreter, eval and macros actually are no problem as long as you're not aiming for speed. But they do become a problem once you start compiling code. Thus the ability of compiling a macro or an eval at execution time seems to me like an elegant solution.
Secondly, scheme's structure allows for easy code analysis that could render a JIT compiler extremely efficient.
In my opinion this could be a killer combination in provoking LISP's regained acceptance.
So what do you think - are scheme and JIT compiling a perfect match?
I have the boxed edition.
I'd probably have chosen the developper installation, but now that you mention it I wonder if I didn't choose something else by mistake. But the thing is, I couldn't even install apache or gdb afterwards from the CDs...
I use Mandrake 7.2 and am generally satisfied with it but I have a few gripes (I'm talking about the full desktop distribution here):
some obviously useful and small tools are lacking : talk, uudecode etc.
some important developper stuff is lacking - most importantly gdb and apache.
I've never had such a failure rate trying to compile some *tar.gz - something is always missing. I'm getting the feeling I can't install anything.
some apps are broken (xchat...) and xemacs has these horrible counter-intuitive icons I cannot begin to fathom why they let the old ones down (by the way if anyone has a solution to this I'd be grateful).
I think what he ment was that ultimately, given enough bandwith, you'd be able to stream audio files from your home server to your portable device - thus obsoleting any physical storage medium for your mobile needs. All you'd need is a big hard drive at home.
Actually this hasn't much to do with Penrose's theory.
His theory is based on Goedel's theorem - basically a Turing Machine machine is weaker that a human mind because we can build formal mathematical proposition based on the Turing Machine's code, that that Turing machine won't be able to prove true or false (it's undecidable in the Turing Machine's formal system) but that we humans can prove true or false thanks to Goedel's theorem.
I personnaly don't subscribe to this theory because I feel his proof is weak around the edges - most importantly a human that proves a theorem doesn't produce the proof in the same sense that a formal system produces a proof. There's always room for a slight error in the human thought process.
On a sidenote, he doesn't give much insight on what he thinks quantum computers might be able to do for AI.
I don't believe peer to peer file sharing must inherently be less efficient than centralised file sharing. I think it's a matter of finding the right protocol. For one thing I think gnutella would benefit from caching and/or indexing. For example, nodes could be organised in supernodes, where each node of a supernode has an index to the files of all the nodes of the supernode.
A few things I'd find interesting from gPulp, Freenet or what have you.
IP tunneling - recreate a whole new web, allocate domain names freely and to first bidder - corporates will never have their say
Interest sheets - by filling out an interest sheet for your node, the software connects tends to connect you to nodes sharing your interests thereby enhacing search retrieval speed on your topics of interest.
Automatic load balancing - if I'm making a heavy use of non freenet services, reduces resources bandwith allocated to freenet. If I'm doing something else (or not using my computer)gives full bandwith to freenet, helps indexing/caching other nodes etc.
On a side note, idealy, I think freenet should concentrate on making a secure encrypted overlay net protocol (IP tunneling, means of logging/using freenet anonymously) while having gPulp and the such concentrate on the file sharing.
Please note, I'm not pretending to have a full grasp of the issues - these are just a few ideas.
If this passes, what this will do is further encourage exchange of video content on the internet, therefore bypassing entirely TV. Maybe there will never be a next generation Tv other than the internet connected PC...
From a moral standpoint I don't see any difference. The main difference lies in the fact that Napster bugs the RIAA because it makes music sharing so easy. Making mix tapes requires a physical medium that costs money. Trading mp3's with Napster costs barely anything and reaches out to thousands of users (as opposed to just your friends) in a very short amount of time.
I think we see way too much napster stories on slashdot these days. Why not do napster quickies or a separate napster theme.
These stories are generating the same endless debate over and over.
Quick summary of the arguments we'll be hearing about:
* Napster is bad cause it's just plain breaking the law.
* Naw every revolution went through its share of law breaking. The outlaws are looked upon as heroes today.
* Napster is just an ISP. The users are responsible.
* Let Napster die! That way a better distributed software will take over.
* When will they learn? The revolution can't be stopped!
* Napster boosts sales (see article).
* They can't shut down Napster because I only download songs I have on CDs.
* They can't shut down Napster because its the way I distribute my songs.
* Come on, you know what Napster is used for! 99% of shared music is pirated!
* Hey I have an idea, why not embed DeCSS in an mp3 and name that file metalica.mp3 and upload that in your shared directory!
I think these debates were interesting, but having them everyday is getting old and I'd really rather have news about gnutella and freenet every now and then.
I think John Searle's argument is interesting but way off the mark.
The man going through the instructions in the book need not be the conscious location of the understanding of chinese for the same reason that neurons are not aware of what they are doing.
Consciousness is not necessarily located somewhere, but if it is, you should at least consider it located in the system formed by the book, the man and everything needed for the interaction to take place.
When you think about it, John Searle has no argument, just clever ways of restating the problem that make you feel he is onto something when in fact he is just stating the obvious : we don't see how consciousness is done.
Do we need these people who make analogies between totally unrelated subjects - draw conclusions without proving the analogy is worthy - and waste our time by stiring everyone's deep fear of the attack by mutant radioactive killer information?
Slashdot Mirror
For example, scheme is hard to compile towards the JVM or standard C because both lack proper tail call implementation (poping the parameters in the stack before calling a function in tail position).
You can of course circumvent this because as you say the target language is Turing complete, but you lose a lot on performance if the target isn't tailored to this particular use.
I'm suprised I've never seen scheme and JIT compiling mentionned together.
First, it seems JIT compiling would elegantly solve the problems linked to eval and macros (and their specific environments). When you program a simple scheme interpreter, eval and macros actually are no problem as long as you're not aiming for speed. But they do become a problem once you start compiling code. Thus the ability of compiling a macro or an eval at execution time seems to me like an elegant solution.
Secondly, scheme's structure allows for easy code analysis that could render a JIT compiler extremely efficient.
In my opinion this could be a killer combination in provoking LISP's regained acceptance.
So what do you think - are scheme and JIT compiling a perfect match?
When will we see a mainstream language implementing multimethods (as in CLOS for example)?
Actually I didn't bother looking on the CDs - I just used their software update program and the programs I mentionned are not in there.
I have the boxed edition. I'd probably have chosen the developper installation, but now that you mention it I wonder if I didn't choose something else by mistake. But the thing is, I couldn't even install apache or gdb afterwards from the CDs...
His theory is based on Goedel's theorem - basically a Turing Machine machine is weaker that a human mind because we can build formal mathematical proposition based on the Turing Machine's code, that that Turing machine won't be able to prove true or false (it's undecidable in the Turing Machine's formal system) but that we humans can prove true or false thanks to Goedel's theorem.
I personnaly don't subscribe to this theory because I feel his proof is weak around the edges - most importantly a human that proves a theorem doesn't produce the proof in the same sense that a formal system produces a proof. There's always room for a slight error in the human thought process.
On a sidenote, he doesn't give much insight on what he thinks quantum computers might be able to do for AI.
5F * 60 = 300F / h
about 40$!
I want shell, gcc, emacs on my PDA.
What happened to russian merced-killer E2K?
Well you're just saying you'll be having high expectations there...
As for freenet, I don't know that much about it.
I don't believe peer to peer file sharing must inherently be less efficient than centralised file sharing. I think it's a matter of finding the right protocol. For one thing I think gnutella would benefit from caching and/or indexing. For example, nodes could be organised in supernodes, where each node of a supernode has an index to the files of all the nodes of the supernode.
IP tunneling - recreate a whole new web, allocate domain names freely and to first bidder - corporates will never have their say
Interest sheets - by filling out an interest sheet for your node, the software connects tends to connect you to nodes sharing your interests thereby enhacing search retrieval speed on your topics of interest.
Automatic load balancing - if I'm making a heavy use of non freenet services, reduces resources bandwith allocated to freenet. If I'm doing something else (or not using my computer)gives full bandwith to freenet, helps indexing/caching other nodes etc.
On a side note, idealy, I think freenet should concentrate on making a secure encrypted overlay net protocol (IP tunneling, means of logging/using freenet anonymously) while having gPulp and the such concentrate on the file sharing.
Please note, I'm not pretending to have a full grasp of the issues - these are just a few ideas.
If this passes, what this will do is further encourage exchange of video content on the internet, therefore bypassing entirely TV. Maybe there will never be a next generation Tv other than the internet connected PC...
From a moral standpoint I don't see any difference. The main difference lies in the fact that Napster bugs the RIAA because it makes music sharing so easy. Making mix tapes requires a physical medium that costs money. Trading mp3's with Napster costs barely anything and reaches out to thousands of users (as opposed to just your friends) in a very short amount of time.
* Napster is bad cause it's just plain breaking the law.
* Naw every revolution went through its share of law breaking. The outlaws are looked upon as heroes today.
* Napster is just an ISP. The users are responsible.
* Let Napster die! That way a better distributed software will take over.
* When will they learn? The revolution can't be stopped!
* Napster boosts sales (see article).
* They can't shut down Napster because I only download songs I have on CDs.
* They can't shut down Napster because its the way I distribute my songs.
* Come on, you know what Napster is used for! 99% of shared music is pirated!
* Hey I have an idea, why not embed DeCSS in an mp3 and name that file metalica.mp3 and upload that in your shared directory!
I think these debates were interesting, but having them everyday is getting old and I'd really rather have news about gnutella and freenet every now and then.
The man going through the instructions in the book need not be the conscious location of the understanding of chinese for the same reason that neurons are not aware of what they are doing.
Consciousness is not necessarily located somewhere, but if it is, you should at least consider it located in the system formed by the book, the man and everything needed for the interaction to take place.
When you think about it, John Searle has no argument, just clever ways of restating the problem that make you feel he is onto something when in fact he is just stating the obvious : we don't see how consciousness is done.
Maybe the answer is there's no consciousness...
The world time isn't US time ... Dresden is in Germany.
Do we need these people who make analogies between totally unrelated subjects - draw conclusions without proving the analogy is worthy - and waste our time by stiring everyone's deep fear of the attack by mutant radioactive killer information?