With a telephone call. And if you exchange two cards in your machine because of a airflow issue which was overheating something Windows now thinks tht you have installed it omn another machine and claims you have to reregister it!
I am working with an app that deals with simple voice recognition (IBM ViaVoice, to be exact). The app works OK, no full dictation, just some very discrete commands.
I was wondering if a Bluetooth headset and a USB dongle would be an adequate mirophone replacement. Voice recognition apps are usually a bit sensitive to mic quality.
Has anyone tries this?
And does any Linux drivers exist that can read the sound stream off a USB dongle?
QNX has its own Microkernel which come back a long way. In the old days the QNX microkernel fit inside the cache on a 486 AFAIK.
Re:Former perl, python, java geek gone to Ruby
on
Ruby 1.8.0 Released
·
· Score: 1
Nor are us Haskell programmer impressed by the reinvention of higher-order functions. However, to see them used in a mainstreamish language IS exciting and should be for you Smalltalkers too. The ideas of Smalltalk and Haskell are finally penetrating the mainstream. Yipee for that.
Re:Former perl, python, java geek gone to Ruby
on
Ruby 1.8.0 Released
·
· Score: 4, Interesting
The major improvement in Ruby over Python and C++ and most OO languages (but not Smalltalk) is the inclusion of code blocks that may be passed in as a parameter to a method and its widespread use throughout the libraries.
For a simple example, look at this:
x = [1,2,3,4]
x.each do |e|
puts e done
This is NOT syntactic sugar. The list class has a each method that takes a piece of code as an argument and executes it for each element in the list. This can do the same as generators now coming into the new Python and is very similar to higher-order functions in functional languages. Basically the do...done part is a parameter (albeit a special one, a method may only have one).
It is difficult and syntactiucally tricky to do this in most OO languages and trivial to add an each method (or whatever you want to call it) to your own classes. It is also extremely powerful way of doing many thing which requires a lot of messiness in other languages.
This high-erorder code blocks now allow you to really do the everything-is-anobject thing because you MUST have code blocks to implement control structures such as if or while as an object. It is also very pervasively used in all collections and for instance in REXML where you can say
x = some XML...
x.each( xpath expression ) do |element|... done
Another example
String.each_regexp( regexp ) do... done
Does the code for each place where the regexp was found in the string. Same sort of thing can be done for databases and so forth.
Ruby also has things like mix-in classes which are usually dissed by OO theoreticians but turn out to be very useful. Mix-in classes were pioneered by some dialect of LISP and are a form of restricted multiple inheritance,. A class can inherit methods from other modules but not instance variables (except for its parent class). This allows you to this (this is pseudocode)
myclass inherits Ord
implement = and other operators in terms of and = (which are supplied by your class). And now your class implements all of these methods. This is again tricky to implement in traditional top-down inheritance trees. Think of Java's interface system, but with default implementations.
This is used to great effect in the collections to implement many things in terms of some primitve operators.
Actually, if there is an official company policy that states that the Tivoli agentmust run it is not exactly "behind your back". They told you they do it and you know it. It is similar to vidoe cameras in shops to nail shoplifters where there is a sign "These premises are guarded by closed circuit camers". Clear enough?
Ah yes, 15 votes for you. Sum types are the greatest. They sumsube enumerators too (as in Color:= Red | Blue | Green) which is another moan in Java.
Sum types also take care of the crap where a numeric type is used to idicate some value or -1 to indicate an absence of value and things like that. Something like x:= Found Int | Notfound.
Thanks for summarizing. It should be a standard feature on ask Slashdot to allow the original submitter to post a summary of what he learned and let this appear at the top at the question.
This would allow the rest of us to sort out the/. "M$ is 3v1l, bad boy!!! bad boy!!!" kneejerkers.
IF you look at the docs for growisofs and the dvd+rwtools package they mention this, yes. They aqlso mention that the kernel needs some work in this area but that some things are underway.
Re:Which DVD drives work under linux
on
DVD Burner Round-up
·
· Score: 2, Informative
Well, I got a DVD drive rcently (a Nec 1300) and it writes both - and +. So I had the same problem.
Here is a quick roundup of my DVD writing adventure.
The +RW tools (growisofs and friends) are very easy to use and work out of the box without any complicated thingies. The program has been called "oddball" by another poster. It is oddball and has the problem of having another interface than CDRecord so most GUI clients do not work with it. XCDRoast for instance does not. However, you could make an ISO with XCDRoast and hand-burn it with growisofs if you want.
The cdrecord-prodvd and dvdrecord, the tools for -RW, uses basically the same command line as cdrecord so the GUI clients are easy to use with it. XCDRoast can use cdrecord-prodvd.
However, however, however. The cdrecord-prodvd recording tool for -R/W is commercial. It does allow you to record full length DVD's but only at 1x speed. At 4x speed it limits you to 1 Gig. Its readdvd feature works very well though. And the demo license that allows this will run out soon. Its not open source either.
There is another GNU cdrecord-prodvd clone called dvdrecord but I could not get it to work properly on Suse 8.2. I have to admit that it could have been the discs though. The NEC is sentive to cheap crap DVD-R's (never buy Princo, even if the salesman who sells you the drive tells you so:)
With -RW you have to blank the disk first, just like CDRW. This did not work with one tool but did with the other but the other tool could record and the one did not. It was a mess. +RW does not need blanking. Pop in the disc and record. Use it again? Pop it in and record again.
In my experience, growisofs, the tool for +rw is easier to use and just works. Nowadays it also can record -RW media but AFAIK it cannot blank a -RW disk.
And it has another sweet feature: There is a kernel driver where you can simply mount your +RW disk as a harddisk and use it as a random-access device.
Of course, DVD-RAM could do this since 5 years ago but the disks are expensive and do not read on normal desktop players. This is an issue for me, because my desktop player can play an ISO full of MP3's and it works with a DVD so I can stuff 50 CD's worth of MP3's on a single disk. Also, few burners handle it, although this is changing.
DVD-RAM is recommended for backups in any case. -RW and +RW both have a maximum amount of times you can write to the disk. RAM does not, you write and rewirte forever. It was designed from the start as a true random-access system.
The maximum write time is a problem with random-access disks because every time you READ a file the Linux fs updates the filesystem to indicate the access time! Do that 1000 times and your +RW disk is wasted. RAM disks on the other hand are designed to be rewirtten pretty much forever.
What really p*ssed me off was that you would think that the stupid drive manufacturers could write those three measly numbers on the hard disk case. But nooooooooo. Maybe the 5 microcents of ink was too much of an extra feature on your $500 20 Meg drive...
Amen to this. I just today had a conversation with a rather nice and intelligent girl at work who got a mail from someone who insisted that she send him a picture "at least 300 DPI". For chrissakes can we just PLEASE get that one thing clear. Maybe we should put it in elementary school education.
DPI IS NOT AN PROPERTY OF A PICTURE. OK? CLEAR? Yes?
The girl understands that now, but the idiot she has to deal with probably does not.
I work as a org which publishes quite a bit of things and we use latex as a back-end typesteer and PDF creation system from all kinds of input.
For instance, our invoicing system produces a large latex file from a database and then uses Tex to crreate the invoices. I also did a long report by using an XSL stylesheet (wchi I could send you) to convert some XML stuff into Latex. This rocks. The first thing my boss said when she pages through the document was "This look eally professional" Latex output really just looks more pleasing to the eye than Word or some other typesetting things.
I also used this XML markup typesetting thing to mark up my GF's PhD thesis and the result was actually quite awesome. There some tweaking to be done but not more than normal. And none of the Microsoft-Worde-screws-you-with-image-placement shit that all her friends had to cope with. (This these was in Immunology so had loads of Microscope slides). In the end we did all the microscope stuff on glossy paper and rest on normal paper anyways so the image palcement, which IS sometimes a pain, did not matter too much. Btw, what the hell is it with Word that places an image so that only the little left part of the corner is actually visible on a page. Why on earth is this the normal, default behaviour??!!
That said, recently I have moved our system to XSL, in particular, FOP. There are two reasons for this
First, Latex sucks with some international characters. We have a system where people can apply for membership on the web and they use all kinds of weirdo character which sort of necessitates unicode. Every once in a while my invoice thingie croaks because some member from some country in Norhern Europe has a funny accent or something on his name and the end stages dies.
Secondly, the, the difference in meta-characters is a pain to use, if someone uses a / in his company name I have to worry. and so forth. This is easier to handle in XML/HTML. The whole business of metacaharacters and the impedance mismatch this causes between stages of a publishing pipeline can be a serious headache. Our system produces text by exporting XML and using the XML to produce Tex. The meaning of a or a \ in an input string can get prety damn confusing.
The third reason is more compelling. Our secretarial staff sometimes needs to update the templates from which we generate EMails. It is mucho easier to do this with a simple subset of HTML (which is what we use) because all of them know the syntax sort of and the other technical guy can help them much easier. I wrote a Java program to process this into XSL:FO and pump it all through FOP while looking up the embedded fields in Lotus Notes. THis works just beautiful, nor problems so far, and quite frankly, the amount of code to manipulate XML in JAva or any other language is muuuuuch more than that to manipulate Tex. A lot of the common metacharacter issues are automatically taken care of, for instance.
Lastly, if you are doing complicated things such as this, Latex's philosophy of "leave the page formatting to me" jsut does not cut it. You can get it right, but it is extremely sensitive to small changes that breaks everything. XSL:FO handles this much better.
I might also add that FOP uses the latex typesetting algorithm. Other XSL:FO rendering tools, RenderX in particular, does not and the output from FOP simply looks better. The output from RenderX looks awful.
I hope that helps. Yes, it is being used in a business setting to produce all kinds of things from papers to invoices with tables and letterheads and whatnot. I might also add that we use these tools because we a a very cost-conscious nonprofit and use Linux for all our servers where a script that happesn to be a Lotus script program outputs fields into a xML file that is used by a PYthon program to convert the output to Tex and then uses Tex to format simply works:)
There is also a XML syntax frontend for Latex AFAIK. If you are really interested I can only reco
The problem with the whole Scifi spaceship dreamy thingie is that the amount of energy required to power a ship, say of Millenium Falcon size, out of the Terran gravity well and then onwards into interstellar space is probably going to be enough to destroy a small city, if not a large one. And if you have that sort of things concentrated in a small ship some nutcase is going to buy one and perform an uncontrolled release.
A jet airliner loaded with fuel has a similar explosive potential as a small tactical nuclear weapon, as New Yorkers found out the hard way, and a spaceship will probably have the potential energy of a very, very large thermonuclear weapon. And if the nice ship is designed to blast off in one shot and zoom into the sunset the powerplant is going to have to be designed to release a large amount of that energy in a short time (unlike nuclear batteries in contemporaty spacecraft which do have a lot of potential but only need to release a small amount of it over a long time).
The long term place for serious energy production lies in antimatter in any case. One gram of antimatter annihilating with matter is enough to completely blow a city-sized hole into the ground, easy. One day the question of whether you want to put that sort of generating capacity into a small, handy penlight sized batter will be a technological problem. Perhaps we should think of the sociological problems before we do that.
But that all lies in the future. A more relevant question is about the here and now. Even today you get quite high energy densities in small devices. Modern Lithium Ion Cellphone batteries made cellphones possible. Your average innocent looking blocky thing inside you cellphone has a thermal and electical fuse inside it to completely shut down the battery if it should ever run out of specs because Lihium Ion batteries can explode. The cellphone makers put this safety mechanism into the batteries because early models blew up next to users heads. The marketing droid referred to this as "discharge with flame". Indeed. What sort of flame would you get from a penlight-sized antimatter batter that some teen geek opened up?
One argument against this is that it depends on how the free energy is delivered. For instance if it is by means of a fusion powerplant driving the electrical grid, you are still limited by the carrying capacity of the network.
Hoever, if you get a situation where someone could get a cute little mini fusion plant in his house which will deliver Gigawatts of energy some other possibilities becomes possible. Read the heat waste argument in the discussion.
Another favourite liberal argument is that there is always the argument that as technology puts more destructive potential into people's hand it also puts more potential to counteract that destruction. Even current technology is quite powerful, one guy in New Zealand (I think) is currently building a demonstration model of a cruise missle. For $5000. There was an article in K5 about this a short while ago. Visionary people like James Gosling are already getting scared by the potantial.
The question is if one guy's experiment will destroy most of humanity before humanity develops a counterattack. Its like a food cycle. If there are too many sharks they eat the fish and then there is not enough fish, the sharks die, and then fish multiply again and then there is more food for the rest of the sharks and cycle continues. Problem is, if that cycle gets off the chart and both species dies off.
And last, "free" is a relative term. "Free" for me means *I* don't have to pay the energy costs of my system. One way of harnessing pretty much free energy is to use a self replicating system where each generation harvests its own energy from the environment, so you, the creator does not have to supply all the energy for the system as a whole to run its course. A Computer virus uses energy paid for by someone else to run on his computer, so for you the virus writer its pretty free. Biological agents are the same thing. The guy who gets infected eats carbohydrates to keep your bioweapon alive.
Until he dies.
The whole "free cheap portable source of energy" problem will create a bad
When Tom Clancy published the Hunt for Red October the US Navy wanted to nail him because they thought he stole some confidential info about their submarine ops.
It turned out that he got all his info from public domain sources. And they could not do much about it. He just knew where to search.
The whole VM issue is a good one, however one also has the whole 80/20 rule. One of the reasons why the Java VM is sometimes slow is because they try to do too much in the VM. For instance, if you use a Python GUI program, it is fast because most of the low-level stuff in Gtk is still written in C. In Java, this goes much deeper and some really low level things are still run in the VM and this makes it slow. Java is more of a 80/20 thing that tries to put 90% of the 20% into the rest of the original 80% which is not always such a good idea.
That said, another solution to the problem is for a program to attach a proof of its ow behaviour with it. This is an interesting avenue which is currently beginning to be explores in academia. Search on Google for TAL and Proof Carrying Code to see some papers.
Basically, a program carries a proof that the computer can verify that is does not cause harm. Java VM's do this in a limited sense, the PCC groups are AFAIK extending some ideas to normal x86 assembler as well as taking it much further. This is also some way of forcing some more serious formal methods on to programs: IF the program cannot prove that it does what is is supposed to do, you do not run it.
It is obviously still several years into the future, but the idea is there and there is quite a bit of research going on at the moment.
Well, I read it because I was bored and my boss is not at work. However, it is AMAZING how many people in the post refer to a Crooke Vane as an example of a Solar Sail when it is explicitly mentioned in the article that it is a counterexample.
Slashdot Mirror
With a telephone call. And if you exchange two cards in your machine because of a airflow issue which was overheating something Windows now thinks tht you have installed it omn another machine and claims you have to reregister it!
Awww cm'on, you can't stop there!
Tell us about the Silence of the Lambs customer with the bodies in the Cellar!
Cascading a failure ona Hydrogen line? Easy. Light a match and stick it in?
Remember the Hindenburg.
I am working with an app that deals with simple voice recognition (IBM ViaVoice, to be exact). The app works OK, no full dictation, just some very discrete commands.
I was wondering if a Bluetooth headset and a USB dongle would be an adequate mirophone replacement. Voice recognition apps are usually a bit sensitive to mic quality.
Has anyone tries this?
And does any Linux drivers exist that can read the sound stream off a USB dongle?
BeOS was written in C++
QNX has its own Microkernel which come back a long way. In the old days the QNX microkernel fit inside the cache on a 486 AFAIK.
Nor are us Haskell programmer impressed by the reinvention of higher-order functions. However, to see them used in a mainstreamish language IS exciting and should be for you Smalltalkers too. The ideas of Smalltalk and Haskell are finally penetrating the mainstream. Yipee for that.
The major improvement in Ruby over Python and C++ and most OO languages (but not Smalltalk) is the inclusion of code blocks that may be passed in as a parameter to a method and its widespread use throughout the libraries.
...
... done
For a simple example, look at this:
x = [1,2,3,4]
x.each do |e|
puts e
done
This is NOT syntactic sugar. The list class has a each method that takes a piece of code as an argument and executes it for each element in the list. This can do the same as generators now coming into the new Python and is very similar to higher-order functions in functional languages.
Basically the do...done part is a parameter (albeit a special one, a method may only have one).
It is difficult and syntactiucally tricky to do this in most OO languages and trivial to add an each method (or whatever you want to call it) to your own classes. It is also extremely powerful way of doing many thing which requires a lot of messiness in other languages.
This high-erorder code blocks now allow you to really do the everything-is-anobject thing because you MUST have code blocks to implement control structures such as if or while as an object. It is also very pervasively used in all collections and for instance in REXML where you can say
x = some XML...
x.each( xpath expression ) do |element|
done
Another example
String.each_regexp( regexp ) do
Does the code for each place where the regexp was found in the string. Same sort of thing can be done for databases and so forth.
Ruby also has things like mix-in classes which are usually dissed by OO theoreticians but turn out to be very useful. Mix-in classes were pioneered by some dialect of LISP and are a form of restricted multiple inheritance,. A class can inherit methods from other modules but not instance variables (except for its parent class). This allows you to this (this is pseudocode)
myclass inherits Ord
implement = and other operators in terms of and = (which are supplied by your class). And now your class implements all of these methods. This is again tricky to implement in traditional top-down inheritance trees. Think of Java's interface system, but with default implementations.
This is used to great effect in the collections to implement many things in terms of some primitve operators.
Actually, if there is an official company policy that states that the Tivoli agentmust run it is not exactly "behind your back". They told you they do it and you know it. It is similar to vidoe cameras in shops to nail shoplifters where there is a sign "These premises are guarded by closed circuit camers". Clear enough?
Call by Name = very bad in a procedural language.
Ah yes, 15 votes for you. Sum types are the greatest. They sumsube enumerators too (as in Color := Red | Blue | Green) which is another moan in Java.
:= Found Int | Notfound.
Sum types also take care of the crap where a numeric type is used to idicate some value or -1 to indicate an absence of value and things like that. Something like x
AFAIk this works with x2vnc at least. The master was x, the second was vnc and that one ran vnc2vnc or something.
Thanks for summarizing. It should be a standard feature on ask Slashdot to allow the original submitter to post a summary of what he learned and let this appear at the top at the question.
/. "M$ is 3v1l, bad boy!!! bad boy!!!" kneejerkers.
This would allow the rest of us to sort out the
IF you look at the docs for growisofs and the dvd+rwtools package they mention this, yes. They aqlso mention that the kernel needs some work in this area but that some things are underway.
Well, I got a DVD drive rcently (a Nec 1300) and it writes both - and +. So I had the same problem.
:)
Here is a quick roundup of my DVD writing adventure.
The +RW tools (growisofs and friends) are very easy to use and work out of the box without any complicated thingies. The program has been called "oddball" by another poster. It is oddball and has the problem of having another interface than CDRecord so most GUI clients do not work with it. XCDRoast for instance does not. However, you could make an ISO with XCDRoast and hand-burn it with growisofs if you want.
The cdrecord-prodvd and dvdrecord, the tools for -RW, uses basically the same command line as cdrecord so the GUI clients are easy to use with it. XCDRoast can use cdrecord-prodvd.
However, however, however. The cdrecord-prodvd recording tool for -R/W is commercial. It does allow you to record full length DVD's but only at 1x speed. At 4x speed it limits you to 1 Gig. Its readdvd feature works very well though. And the demo license that allows this will run out soon. Its not open source either.
There is another GNU cdrecord-prodvd clone called dvdrecord but I could not get it to work properly on Suse 8.2. I have to admit that it could have been the discs though. The NEC is sentive to cheap crap DVD-R's (never buy Princo, even if the salesman who sells you the drive tells you so
With -RW you have to blank the disk first, just like CDRW. This did not work with one tool but did with the other but the other tool could record and the one did not. It was a mess. +RW does not need blanking. Pop in the disc and record. Use it again? Pop it in and record again.
In my experience, growisofs, the tool for +rw is easier to use and just works. Nowadays it also can record -RW media but AFAIK it cannot blank a -RW disk.
And it has another sweet feature: There is a kernel driver where you can simply mount your +RW disk as a harddisk and use it as a random-access device.
Of course, DVD-RAM could do this since 5 years ago but the disks are expensive and do not read on normal desktop players. This is an issue for me, because my desktop player can play an ISO full of MP3's and it works with a DVD so I can stuff 50 CD's worth of MP3's on a single disk. Also, few burners handle it, although this is changing.
DVD-RAM is recommended for backups in any case. -RW and +RW both have a maximum amount of times you can write to the disk. RAM does not, you write and rewirte forever. It was designed from the start as a true random-access system.
The maximum write time is a problem with random-access disks because every time you READ a file the Linux fs updates the filesystem to indicate the access time! Do that 1000 times and your +RW disk is wasted. RAM disks on the other hand are designed to be rewirtten pretty much forever.
Oh yeah the bad old days.
What really p*ssed me off was that you would think that the stupid drive manufacturers could write those three measly numbers on the hard disk case. But nooooooooo. Maybe the 5 microcents of ink was too much of an extra feature on your $500 20 Meg drive...
Yes he could. And he did not specify the final print size...
Amen to this. I just today had a conversation with a rather nice and intelligent girl at work who got a mail from someone who insisted that she send him a picture "at least 300 DPI". For chrissakes can we just PLEASE get that one thing clear. Maybe we should put it in elementary school education.
DPI IS NOT AN PROPERTY OF A PICTURE. OK? CLEAR? Yes?
The girl understands that now, but the idiot she has to deal with probably does not.
I work as a org which publishes quite a bit of things and we use latex as a back-end typesteer and PDF creation system from all kinds of input.
:)
For instance, our invoicing system produces a large latex file from a database and then uses Tex to crreate the invoices. I also did a long report by using an XSL stylesheet (wchi I could send you) to convert some XML stuff into Latex. This rocks. The first thing my boss said when she pages through the document was "This look eally professional" Latex output really just looks more pleasing to the eye than Word or some other typesetting things.
I also used this XML markup typesetting thing to mark up my GF's PhD thesis and the result was actually quite awesome. There some tweaking to be done but not more than normal. And none of the Microsoft-Worde-screws-you-with-image-placement shit that all her friends had to cope with. (This these was in Immunology so had loads of Microscope slides). In the end we did all the microscope stuff on glossy paper and rest on normal paper anyways so the image palcement, which IS sometimes a pain, did not matter too much. Btw, what the hell is it with Word that places an image so that only the little left part of the corner is actually visible on a page. Why on earth is this the normal, default behaviour??!!
That said, recently I have moved our system to XSL, in particular, FOP. There are two reasons for this
First, Latex sucks with some international characters. We have a system where people can apply for membership on the web and they use all kinds of weirdo character which sort of necessitates unicode. Every once in a while my invoice thingie croaks because some member from some country in Norhern Europe has a funny accent or something on his name and the end stages dies.
Secondly, the, the difference in meta-characters is a pain to use, if someone uses a / in his company name I have to worry. and so forth. This is easier to handle in XML/HTML. The whole business of metacaharacters and the impedance mismatch this causes between stages of a publishing pipeline can be a serious headache. Our system produces text by exporting XML and using the XML to produce Tex. The meaning of a or a \ in an input string can get prety damn confusing.
The third reason is more compelling. Our secretarial staff sometimes needs to update the templates from which we generate EMails. It is mucho easier to do this with a simple subset of HTML (which is what we use) because all of them know the syntax sort of and the other technical guy can help them much easier. I wrote a Java program to process this into XSL:FO and pump it all through FOP while looking up the embedded fields in Lotus Notes. THis works just beautiful, nor problems so far, and quite frankly, the amount of code to manipulate XML in JAva or any other language is muuuuuch more than that to manipulate Tex. A lot of the common metacharacter issues are automatically taken care of, for instance.
Lastly, if you are doing complicated things such as this, Latex's philosophy of "leave the page formatting to me" jsut does not cut it. You can get it right, but it is extremely sensitive to small changes that breaks everything. XSL:FO handles this much better.
I might also add that FOP uses the latex typesetting algorithm. Other XSL:FO rendering tools, RenderX in particular, does not and the output from FOP simply looks better. The output from RenderX looks awful.
I hope that helps. Yes, it is being used in a business setting to produce all kinds of things from papers to invoices with tables and letterheads and whatnot. I might also add that we use these tools because we a a very cost-conscious nonprofit and use Linux for all our servers where a script that happesn to be a Lotus script program outputs fields into a xML file that is used by a PYthon program to convert the output to Tex and then uses Tex to format simply works
There is also a XML syntax frontend for Latex AFAIK. If you are really interested I can only reco
The real question is, do you WANT to have your child read "There's adventure in Atomic Energy" followed by "There's adventure in Rockets"??!!
The problem with the whole Scifi spaceship dreamy thingie is that the amount of energy required to power a ship, say of Millenium Falcon size, out of the Terran gravity well and then onwards into interstellar space is probably going to be enough to destroy a small city, if not a large one. And if you have that sort of things concentrated in a small ship some nutcase is going to buy one and perform an uncontrolled release.
A jet airliner loaded with fuel has a similar explosive potential as a small tactical nuclear weapon, as New Yorkers found out the hard way, and a spaceship will probably have the potential energy of a very, very large thermonuclear weapon. And if the nice ship is designed to blast off in one shot and zoom into the sunset the powerplant is going to have to be designed to release a large amount of that energy in a short time (unlike nuclear batteries in contemporaty spacecraft which do have a lot of potential but only need to release a small amount of it over a long time).
The long term place for serious energy production lies in antimatter in any case. One gram of antimatter annihilating with matter is enough to completely blow a city-sized hole into the ground, easy. One day the question of whether you want to put that sort of generating capacity into a small, handy penlight sized batter will be a technological problem. Perhaps we should think of the sociological problems before we do that.
But that all lies in the future. A more relevant question is about the here and now. Even today you get quite high energy densities in small devices. Modern Lithium Ion Cellphone batteries made cellphones possible. Your average innocent looking blocky thing inside you cellphone has a thermal and electical fuse inside it to completely shut down the battery if it should ever run out of specs because Lihium Ion batteries can explode. The cellphone makers put this safety mechanism into the batteries because early models blew up next to users heads. The marketing droid referred to this as "discharge with flame". Indeed. What sort of flame would you get from a penlight-sized antimatter batter that some teen geek opened up?
One argument against this is that it depends on how the free energy is delivered. For instance if it is by means of a fusion powerplant driving the electrical grid, you are still limited by the carrying capacity of the network.
Hoever, if you get a situation where someone could get a cute little mini fusion plant in his house which will deliver Gigawatts of energy some other possibilities becomes possible. Read the heat waste argument in the discussion.
Another favourite liberal argument is that there is always the argument that as technology puts more destructive potential into people's hand it also puts more potential to counteract that destruction. Even current technology is quite powerful, one guy in New Zealand (I think) is currently building a demonstration model of a cruise missle. For $5000. There was an article in K5 about this a short while ago. Visionary people like James Gosling are already getting scared by the potantial.
The question is if one guy's experiment will destroy most of humanity before humanity develops a counterattack. Its like a food cycle. If there are too many sharks they eat the fish and then there is not enough fish, the sharks die, and then fish multiply again and then there is more food for the rest of the sharks and cycle continues. Problem is, if that cycle gets off the chart and both species dies off.
And last, "free" is a relative term. "Free" for me means *I* don't have to pay the energy costs of my system. One way of harnessing pretty much free energy is to use a self replicating system where each generation harvests its own energy from the environment, so you, the creator does not have to supply all the energy for the system as a whole to run its course. A Computer virus uses energy paid for by someone else to run on his computer, so for you the virus writer its pretty free. Biological agents are the same thing. The guy who gets infected eats carbohydrates to keep your bioweapon alive.
Until he dies.
The whole "free cheap portable source of energy" problem will create a bad
When Tom Clancy published the Hunt for Red October the US Navy wanted to nail him because they thought he stole some confidential info about their submarine ops.
It turned out that he got all his info from public domain sources. And they could not do much about it. He just knew where to search.
The whole VM issue is a good one, however one also has the whole 80/20 rule. One of the reasons why the Java VM is sometimes slow is because they try to do too much in the VM. For instance, if you use a Python GUI program, it is fast because most of the low-level stuff in Gtk is still written in C. In Java, this goes much deeper and some really low level things are still run in the VM and this makes it slow. Java is more of a 80/20 thing that tries to put 90% of the 20% into the rest of the original 80% which is not always such a good idea.
That said, another solution to the problem is for a program to attach a proof of its ow behaviour with it. This is an interesting avenue which is currently beginning to be explores in academia. Search on Google for TAL and Proof Carrying Code to see some papers.
Basically, a program carries a proof that the computer can verify that is does not cause harm. Java VM's do this in a limited sense, the PCC groups are AFAIK extending some ideas to normal x86 assembler as well as taking it much further. This is also some way of forcing some more serious formal methods on to programs: IF the program cannot prove that it does what is is supposed to do, you do not run it.
It is obviously still several years into the future, but the idea is there and there is quite a bit of research going on at the moment.
Google turned up a demo version. After struggling with the download it turns out to be in... Japanese!
Anyone knows if there is an English/German/Dutch version somewhere??!!
What about feeding the homeless to the hungry? That better?
Well, I read it because I was bored and my boss is not at work. However, it is AMAZING how many people in the post refer to a Crooke Vane as an example of a Solar Sail when it is explicitly mentioned in the article that it is a counterexample.