One thing I think that few people realise is that a lot of the features that MS have in their OS's and office software come out of sales conversations with CIO's in Fortune 500 companies, and the teams that are working with them. Things that the average slashdotter hates (I'm guessing here!) such as the complex reviewing functions in word or the start bar in windows, or the difficulty in finding the f'n command line and using it when it is found, are things that corporates love. Have a good look at what all the civilians in your organization are doing with their machines.
Other features are tactical by MS to block third parties (like the environment editors in windows 2000 - why so bad... so bad stop you have good classpath edit!)
Of course there are things that are basically unfathomable about MS software as well...
I may be wrong here, but I have a strong impression that there is a very large input from IBM into Apache? I think that this may be a blocking tactic to drain revenue from potential future competitors thus protecting Big Blues long term position.
The critical movement in technology for GPS systems is the fundamental improvement of the models for calculating positions on earth from triangulations. Relative orbital positions are influenced by the relative postions of the oceans, geology and the positions of Jupiter, Sol and other planets. This is nothing to do with what is in the network itself.
The Russians run a GPS network of their own. It is called GLONASS and has been operational for 20 years, and is currently as well maintained, or better maintained (given US launcher problems) than GPS.
The Chinese have no real concerns about the GPS signal due to the fact that if they decide to stop trading with the US then everyone in the middle of America starves as the lights go out.
What is "coding in Java" apart from specifying what things are supposed to do? Well...
The problem of Cobol to Java converters (and they've been around for some time) is that the expressive power of Cobol, while sufficient for encoding a program does not contain information like, for example, object decompositions that allow its comprehension for reuse. Strangely, langauges like Java, C++ and Smalltalk often provide notation for encoding these knowledge items.
It's not talk; they tested a hypersonic manuvering reentry craft for warheads a couple of months ago; basically this is a system that let's the warhead dodge around like a balloon with another ballon full of water inside it (ever try to catch one like that?) I don't know anyone who has a fucking clue how they did that, although one old fellow who I did talk to about it said:
"ohhhh, that's interesting, how did they do that, that means that they must have solved the..." catches self, looks shifty, "well, that's very interesting..." looks more shifty.
And do you know what I saw this all on the news in the UK, and guess what, there is F*all ^F*all on google; yet see
http://cndyorks.gn.apc.org/yspace/articles/bmd/r us sian_hypersonic_missile.htm
from search.msn.com no less. Well F! me.
And that, old chap is not all. The Russians have operated an ABM system to protect Moscow and their launch pads for more than 20 years. They are allow to do this under the ABM treaty, as is the US (capitol city + launcher protection is ok) The difference between Russia and the US is that there are very few H-Bomb targets in Russia, whereas LA, Chicago, Dallas, NYC, SF, Seattle... well we could go on.
Now; why do the russians possess these things, while the USA hasn't until just now? Because until just a few years ago the Russians were the only barking mad ones. Now we're all mad.
Wibble, wibble, wibble.
noooooooo......
Repeat after me. This is a business that no one can win at. This is a business that no one can win at. This...
Isn't anyone else surprised that the critters have never evolved new and better fighting styles? I mean a human in a forest could within an hour create enough weapons to fight off a single baboon or two (well, maybe not a chair-bound worker from America but you get my point:-). Why haven't these critters learned anything from humans?
Because they are Baboons, and Baboons are not real bright.
Ok flippancy aside; no one knows is the real answer. But one theory is that the ability to figure out how to crack open long bones from scavenged out carcasses on the savannah provided the fuel for an evolutionary upward spiral; more protein = more smarts = more breeding = more smarts = more protein...
Another theory is that Humans were aquatic adapted apes (we have a lack of body hair, subcutaneous fat layer, fat distributions for floating face up, swimming ability, dive reflex all of which could be argued to be aquatic adaptations). We started eating fish and shellfish and promptly acquired smarts. Interestingly evidence indicates that fish oils can be given to children to boost their brain power, and can be given to prisoners to boost their problem solving and social skills! Also interesting is the fact that humans seem to only like raw fish and shell fish amongst raw meats Kobe Beef aside??? Unfortunately there is not a shred of archaeological evidence to support these ideas, and some people argue that the so called aquatic adaptations are't really aquatic in nature and just a co-incidence (still remember the baby on the front of Nevermind - it looked pretty adapted!)
Finally the most prosaic theory is that standing upright enabled superior brain cooling (I kid you not) and thus enabled big brains for all.
But as I say, anyone who knows says nothing - as they were there, and they are dead!
I guess this is a key fissure in computing at the moment.
Tools for designing workflows, like ARIS, don't go together with tools for designing programs, like Rose. At the moment there is a big movement to intergrate service, process and computation in the GRID and Web-Services communities. These discussions are mirrored in the Agent community and OMG's MDA. Languages like BPML4WS do not sit well with activity diagrams and class diagrams.
I think this is because in the workflow world one type of event (a transition) is assumed, where as in the OO world many types of events are possible, but the language doesn't model them. This is a break in the notation that we have. Does anyone have a good solution?
well - you can! if you assume someing is not not true and then extrapolate a contradiction from that then you know that it's not not not true and so it must be not true...
Heres the issue. Blade based servers are cheap and easy to run, they don't take much space and power and they have been developed in an endless quest to reduce costs.
Utility computing is new technology. It doesn't make the need for computing capacity and investment go away, but it has to compete with mature solutions like blades and corporate data centers in order to win.
So the problem is this, the market is highly competive as it is, and the winners in the new world of the GRID will not be the hardware manufactures, who by definition will be selling less (mission critical) kit. Yet these very companies (HP,IBM) are the ones pushing the vision... why so?
My guess is that this is a not-a-zero-sum play. IBM hopes that by enabling another layer of applications with computing on demand (the business case for those marginal uses of computing just got easier to write..) it will be able to addict enterprises to them. The idea being that what is an meaningless luxury that can't get a business case signed off on it to save it's life suddenly becomes mission critical.
Will it work? I think so. I think that the ruthless darwinianism of corporate life (stop laughing fatso, it's your job off to India next) will winnow out loser applications and the ones that remain will successfully free up investment capital to be ploughed back into big blues bottom line.
Having said that there are quite a few little problems with GRID computing still to be resolved...
I think the odds must be good - their is a method that uses sighting of a species to derive a date at which the species actually went extinct, normally this is a long time after the final sighting. The idea is that as sightings get rarer the chances of another animal turning up get lower, so if Dr Who episodes are running at one every 6 years I think we can expect a couple more to pop up!
- Beagle was much lighter than Spirit and so did not include retro rockets in the design. The Spirit system required retros because the weight would cause the impact stress to exceed the maximum stress that airbags can endure.
- Beagle did not have decent communication, again due to weight. This time though they didn't put this feature in in order to save weight and cost. The parachute was deployed after re-entry (if it was deployed at all) and so there was no contact with the craft at this time.
I saw an interesting documentary about Spirit over christmas. On it they showed that the Nasa team had problems with both the airbags tearing and the chute failing to open in wind tunnel tests. I think that the Nasa budget enabled extensive testing and fixing of these problems, while the shoestring Beagle effort was able to do only limited testing. Event then I believe that the airbags were a major issue for Beagle.
Of course Beagle is a success in relative terms, the probe was built, launched, taken to Mars orbit, and released. I wish I had ever been involved in a project that did such a string of hard things so well. Sadly, there was at least one thing that doesn't seem to have worked. I hope that the little bugger wakes up in the next few months and proves us all wrong though. Roll on Beagle 3!!!!
Ok VC's are evil, we hate them, they stole our company, they were nasty to us when we pitched, they were nasty when the company ran, they were nasty when the company folded...
VC's are mostly: clever; honest; brutal; gamblers. Ok, it's not a great set of characteristics, but they're like sharks, not fun to play with in the pool, but noble beasts by their own lights. Don't hate them for what they are!!! All the VC's I have dealt with have been straight with me "it's our money, it's our company, you will do what we say, if it goes wrong we blame you, your ideas and you lose." not nice, but honest.
Most people who take VC funding do it for one of three reasons:
they have to or they will be bust in a month; this means that they have already lost their company and just hope for a few more paychecks and a reasonable pay off in the best case
they are in a strong position, they want to expand, they are quick hard and clever and have a strong enough cash position not to have to take a follow up deal (that will kill their stake)
they read stories about how dumb VC's are just handing out money for free. They think that the VC's are dumb and that the VC's will get ripped off, by them... oh dear, oh dear...
The reality is that venture funds have to make money and the successes (Amazon, eBay, yahoo) of the last tech wave have paid for the failures. I would love to see the sums though but I imagine that the investment in Amazon has paid for hundereds of misfires..... My point is that driving companies to ipo or a trade sale (much preferred) is not evil; who knows what the health of a company after 5 years will be in any case?
Because copyright law is well known and determined by the Berne convention.
Here are the pertinant facts
You cannot loose your copyright on something that you wrote. It is yours for life unless you assign it to someone
You can ceed it to someone else under a license, but your protection under copyright prevents them from every taking that license off you.
When you work for someone the copyright on the work you produce during the time that they pay you belongs to them, and they can do what they want with it. It is not clear what determines the copyright of items created for a company.
Copyright applies to code, text, music and video.
Copyright has a stronger status than a patent in law because it is easier to prove a violation of it (here is the *copy* that you have made instead of here is the *idea* you used) But items that are copyrighted by someone can be protected by a patent, and licensed items can be protected by patents. This is the killer for Linux and will be how people get it if they every get it, because if someone has a patent on a GPL'ed item they will be able to enforce that patent on derivitive works that are not covered by the GPL and it is argueable that a rewritten class is separated from the initial license because the copyright has now passed to the author of the rewrite (who can grant a license to the copyright, but not to the patent)
Democracy can only last until the people realise that they can vote themselves bounty from the state. This is not a good thing, people vote for the guy who says he'll be best for them; it's game theory really. The idiots who vote for the guy who'll be best for everyone always get screwed...
In the future people will shake their heads about this. They won't laugh though.
"The inconsistencies between what the language allows and what the standard library actually does bother me. If operator overloading is so bad, why does the String class do it?"
Yo! who cares, yeah, sooon you be decompiling string, yo! you neeeeeed work to do boooooy.
"Interfaces get around part of the lack of multiple inheritance, but I'd like to be able to reuse common code in ways besides inheritance. Mixins that don't require inheritance would be a nice touch."
Dat is why we have AspectJ my brother. Shit, if you need it, use it, ain't no one judging you. But you know, you be bringing that to my house and I'll put a cap in your ass. Straight.
"The libraries and the interpreter aren't cleanly separated. There are ways (involving decompilers), for example, to get regex support in 1.3, but I'd prefer to upgrade the standard library and the interpreter separately sometimes, rather than in one big chunk."
Ok, you have a point here. This badness can do some harm.
"I like the idea of checked exceptions in some situations, but forcing every method to deal with (catching or throwing) all exceptions that its child calls or may call can be tedious. I'd rather be able to ignore an exception and let it propagate upwards. Sometimes, I'd rather not worry about exceptions at all."
My man, you know shit.
public Shit someMfkinMeth (Shit shitbooooy) throws throwable {
It's always interesting how people argue about this without reference to what we know about how things are in our world.
It is the case that states of afairs are represented by sets of symbols. It is the case that propositions, or truths are the result of states of afairs. Now, an inference is a relation between two propositions, and its nature is only understood from the nature of the two propositions itself. All rules of inference are irrelevant, because the nature of the inferences that are true is determined a-priori, for example the fact that my chair is on the floor is not derived from an understanding of the laws of gravity, but rather from the arrangement and nature of my chair and the floor. No elementary proposition can be deduced from another one. There is no way of inferencing from the existance of one sitation to the existance of another different situation. There is no causality. There simply is, an is.
Now, freedom of will is derived from the impossibility of knowing actions that lie in the future, while Strong AI is the belief that processes that manipulate physical devices of some sort are capable of actually becoming intelligent; in short that inferences between situations can be made. This may be true, but the only possible class of these inferences is the class of inferences that are true a-priori; thus the only inferences that an AI can make will be those that it can be predicted (possibly it will be very difficult) to make. Thus a machine of any type can appear to be intelligent, but it cannot have free will, and therefore it will be incapable of changing the world in which it acts; beyond the changes that could be directly infered from its introduction into that world. For example, a asteroid striking London would not change global politics, because the asteroid was always going to strike London, however the fact that people built London where it is and then it has become significant to other people is the thing that makes it's (putative) destruction significant
There is a difference, and if you wrap your silly little mid western mind around a half decent book like Tractatus Logico-Philosophicus you will figure it out! My point?
Don't worry, this ain't going to happen, because it simply can't happen in this reality.
I would like to see a system where you were able to have two mail boxes - one that was public and caught all unsolicited mail. This could have a list of authorised users, and could forward mail from them to you. Non-authorised users could buy a stamp as suggested here, and then the public mail box could forward it on.
This would mean that people could pay to contact you, for example to change address or get in touch. Your friends and close collegues could use email as in the good old days.
Problems with swarm implementations
on
Swarm Intelligence
·
· Score: 2, Interesting
Sure some of the demos are excellent. The problem is these implementations require just as much (at least) work as engineered solutions; it's just a different point of effort! Basically you have to discover how to make the self organising system organise itself to solve your problem, otherwise - it'll just sit and look at you. The trouble with this is that you don't know if it will work, until it works - a big risk for a commercial development project.
RMS's goals are fine, I can see some of his point (but remember, you aren't allowed to share my house or salary, and I'm going to try and stop you sharing my code as well!). The problem is that free software with proprietary software is an economic good.
Some software is literally not worth marketing, it can be reverse engineered, can't support a business model because it is only useful for 100 people world wide or is background value - say worth 50cents a user a year. This is not to say that it is worthless, just not worth a marketing budget and a business. Giving away software like this is economically good, it saves many people time. Allowing those people to use it in commercial products is ecomonically better; rapid development of high quality stuff is enabled.
What we need to do is work out a system for moving some of the money back to the original contributers so that things are fair, and people are encouraged to do this. Think : you spend three wet weekends doing something, someone takes it and makes $millions, you never see a cent: fair? No! Should you get the $millions? NO! But you should see something for your skills.
Slashdot Mirror
Good points.
One thing I think that few people realise is that a lot of the features that MS have in their OS's and office software come out of sales conversations with CIO's in Fortune 500 companies, and the teams that are working with them. Things that the average slashdotter hates (I'm guessing here!) such as the complex reviewing functions in word or the start bar in windows, or the difficulty in finding the f'n command line and using it when it is found, are things that corporates love. Have a good look at what all the civilians in your organization are doing with their machines.
Other features are tactical by MS to block third parties (like the environment editors in windows 2000 - why so bad... so bad stop you have good classpath edit!)
Of course there are things that are basically unfathomable about MS software as well...
I may be wrong here, but I have a strong impression that there is a very large input from IBM into Apache? I think that this may be a blocking tactic to drain revenue from potential future competitors thus protecting Big Blues long term position.
The critical movement in technology for GPS systems is the fundamental improvement of the models for calculating positions on earth from triangulations. Relative orbital positions are influenced by the relative postions of the oceans, geology and the positions of Jupiter, Sol and other planets. This is nothing to do with what is in the network itself.
The Russians run a GPS network of their own. It is called GLONASS and has been operational for 20 years, and is currently as well maintained, or better maintained (given US launcher problems) than GPS.
The Chinese have no real concerns about the GPS signal due to the fact that if they decide to stop trading with the US then everyone in the middle of America starves as the lights go out.
What is "coding in Java" apart from specifying what things are supposed to do? Well...
The problem of Cobol to Java converters (and they've been around for some time) is that the expressive power of Cobol, while sufficient for encoding a program does not contain information like, for example, object decompositions that allow its comprehension for reuse. Strangely, langauges like Java, C++ and Smalltalk often provide notation for encoding these knowledge items.
http://msnbc.msn.com/id/4317546
It's not talk; they tested a hypersonic manuvering reentry craft for warheads a couple of months ago; basically this is a system that let's the warhead dodge around like a balloon with another ballon full of water inside it (ever try to catch one like that?) I don't know anyone who has a fucking clue how they did that, although one old fellow who I did talk to about it said :
r us sian_hypersonic_missile.htm
...
"ohhhh, that's interesting, how did they do that, that means that they must have solved the..." catches self, looks shifty, "well, that's very interesting..." looks more shifty.
And do you know what I saw this all on the news in the UK, and guess what, there is F*all ^F*all on google; yet see
http://cndyorks.gn.apc.org/yspace/articles/bmd/
from search.msn.com no less. Well F! me.
And that, old chap is not all. The Russians have operated an ABM system to protect Moscow and their launch pads for more than 20 years. They are allow to do this under the ABM treaty, as is the US (capitol city + launcher protection is ok) The difference between Russia and the US is that there are very few H-Bomb targets in Russia, whereas LA, Chicago, Dallas, NYC, SF, Seattle... well we could go on.
Now; why do the russians possess these things, while the USA hasn't until just now? Because until just a few years ago the Russians were the only barking mad ones. Now we're all mad.
Wibble, wibble, wibble.
noooooooo......
Repeat after me. This is a business that no one can win at. This is a business that no one can win at. This
And this morning they reported a half year profit of £546 million - that equates to £1000 million per year or $1500 million.
m ?ArticleID=18949ba8-e80d-40fc-8e0e-649e8a2a8388
See http://www.btplc.com/News/Articles/Showarticle.cf
The sweetest thing is that Wal-Mart are one of the new wave of companies that do thier programming in house, no offshoring, and no COTS software.
Isn't anyone else surprised that the critters have never evolved new and better fighting styles? I mean a human in a forest could within an hour create enough weapons to fight off a single baboon or two (well, maybe not a chair-bound worker from America but you get my point :-). Why haven't these critters learned anything from humans?
Because they are Baboons, and Baboons are not real bright.
Ok flippancy aside; no one knows is the real answer. But one theory is that the ability to figure out how to crack open long bones from scavenged out carcasses on the savannah provided the fuel for an evolutionary upward spiral; more protein = more smarts = more breeding = more smarts = more protein...
Another theory is that Humans were aquatic adapted apes (we have a lack of body hair, subcutaneous fat layer, fat distributions for floating face up, swimming ability, dive reflex all of which could be argued to be aquatic adaptations). We started eating fish and shellfish and promptly acquired smarts. Interestingly evidence indicates that fish oils can be given to children to boost their brain power, and can be given to prisoners to boost their problem solving and social skills! Also interesting is the fact that humans seem to only like raw fish and shell fish amongst raw meats Kobe Beef aside??? Unfortunately there is not a shred of archaeological evidence to support these ideas, and some people argue that the so called aquatic adaptations are't really aquatic in nature and just a co-incidence (still remember the baby on the front of Nevermind - it looked pretty adapted!)
Finally the most prosaic theory is that standing upright enabled superior brain cooling (I kid you not) and thus enabled big brains for all.
But as I say, anyone who knows says nothing - as they were there, and they are dead!
I guess this is a key fissure in computing at the moment.
Tools for designing workflows, like ARIS, don't go together with tools for designing programs, like Rose. At the moment there is a big movement to intergrate service, process and computation in the GRID and Web-Services communities. These discussions are mirrored in the Agent community and OMG's MDA. Languages like BPML4WS do not sit well with activity diagrams and class diagrams.
I think this is because in the workflow world one type of event (a transition) is assumed, where as in the OO world many types of events are possible, but the language doesn't model them. This is a break in the notation that we have. Does anyone have a good solution?
well - you can! if you assume someing is not not true and then extrapolate a contradiction from that then you know that it's not not not true and so it must be not true...
or false!
Heres the issue. Blade based servers are cheap and easy to run, they don't take much space and power and they have been developed in an endless quest to reduce costs.
Utility computing is new technology. It doesn't make the need for computing capacity and investment go away, but it has to compete with mature solutions like blades and corporate data centers in order to win.
So the problem is this, the market is highly competive as it is, and the winners in the new world of the GRID will not be the hardware manufactures, who by definition will be selling less (mission critical) kit. Yet these very companies (HP,IBM) are the ones pushing the vision... why so?
My guess is that this is a not-a-zero-sum play. IBM hopes that by enabling another layer of applications with computing on demand (the business case for those marginal uses of computing just got easier to write..) it will be able to addict enterprises to them. The idea being that what is an meaningless luxury that can't get a business case signed off on it to save it's life suddenly becomes mission critical.
Will it work? I think so. I think that the ruthless darwinianism of corporate life (stop laughing fatso, it's your job off to India next) will winnow out loser applications and the ones that remain will successfully free up investment capital to be ploughed back into big blues bottom line.
Having said that there are quite a few little problems with GRID computing still to be resolved...
ok, ok; what do you call it though?
Mars insertion just sounds too bad!!
I think the odds must be good - their is a method that uses sighting of a species to derive a date at which the species actually went extinct, normally this is a long time after the final sighting. The idea is that as sightings get rarer the chances of another animal turning up get lower, so if Dr Who episodes are running at one every 6 years I think we can expect a couple more to pop up!
- Beagle was much lighter than Spirit and so did not include retro rockets in the design. The Spirit system required retros because the weight would cause the impact stress to exceed the maximum stress that airbags can endure.
- Beagle did not have decent communication, again due to weight. This time though they didn't put this feature in in order to save weight and cost. The parachute was deployed after re-entry (if it was deployed at all) and so there was no contact with the craft at this time.
I saw an interesting documentary about Spirit over christmas. On it they showed that the Nasa team had problems with both the airbags tearing and the chute failing to open in wind tunnel tests. I think that the Nasa budget enabled extensive testing and fixing of these problems, while the shoestring Beagle effort was able to do only limited testing. Event then I believe that the airbags were a major issue for Beagle.
Of course Beagle is a success in relative terms, the probe was built, launched, taken to Mars orbit, and released. I wish I had ever been involved in a project that did such a string of hard things so well. Sadly, there was at least one thing that doesn't seem to have worked. I hope that the little bugger wakes up in the next few months and proves us all wrong though. Roll on Beagle 3!!!!
Ok VC's are evil, we hate them, they stole our company, they were nasty to us when we pitched, they were nasty when the company ran, they were nasty when the company folded...
VC's are mostly: clever; honest; brutal; gamblers. Ok, it's not a great set of characteristics, but they're like sharks, not fun to play with in the pool, but noble beasts by their own lights. Don't hate them for what they are!!! All the VC's I have dealt with have been straight with me "it's our money, it's our company, you will do what we say, if it goes wrong we blame you, your ideas and you lose." not nice, but honest.
Most people who take VC funding do it for one of three reasons:
they have to or they will be bust in a month; this means that they have already lost their company and just hope for a few more paychecks and a reasonable pay off in the best case
they are in a strong position, they want to expand, they are quick hard and clever and have a strong enough cash position not to have to take a follow up deal (that will kill their stake)
they read stories about how dumb VC's are just handing out money for free. They think that the VC's are dumb and that the VC's will get ripped off, by them... oh dear, oh dear...
The reality is that venture funds have to make money and the successes (Amazon, eBay, yahoo) of the last tech wave have paid for the failures. I would love to see the sums though but I imagine that the investment in Amazon has paid for hundereds of misfires..... My point is that driving companies to ipo or a trade sale (much preferred) is not evil; who knows what the health of a company after 5 years will be in any case?
Scotland is though, and so are the famous urban centers of Sweden.
Because copyright law is well known and determined by the Berne convention.
Here are the pertinant facts
You cannot loose your copyright on something that you wrote. It is yours for life unless you assign it to someone
You can ceed it to someone else under a license, but your protection under copyright prevents them from every taking that license off you.
When you work for someone the copyright on the work you produce during the time that they pay you belongs to them, and they can do what they want with it. It is not clear what determines the copyright of items created for a company.
Copyright applies to code, text, music and video.
Copyright has a stronger status than a patent in law because it is easier to prove a violation of it (here is the *copy* that you have made instead of here is the *idea* you used) But items that are copyrighted by someone can be protected by a patent, and licensed items can be protected by patents. This is the killer for Linux and will be how people get it if they every get it, because if someone has a patent on a GPL'ed item they will be able to enforce that patent on derivitive works that are not covered by the GPL and it is argueable that a rewritten class is separated from the initial license because the copyright has now passed to the author of the rewrite (who can grant a license to the copyright, but not to the patent)
In the future people will shake their heads about this. They won't laugh though.
"The inconsistencies between what the language allows and what the standard library actually does bother me. If operator overloading is so bad, why does the String class do it?"
Yo! who cares, yeah, sooon you be decompiling string, yo! you neeeeeed work to do boooooy.
"Interfaces get around part of the lack of multiple inheritance, but I'd like to be able to reuse common code in ways besides inheritance. Mixins that don't require inheritance would be a nice touch."
Dat is why we have AspectJ my brother. Shit, if you need it, use it, ain't no one judging you. But you know, you be bringing that to my house and I'll put a cap in your ass. Straight.
"The libraries and the interpreter aren't cleanly separated. There are ways (involving decompilers), for example, to get regex support in 1.3, but I'd prefer to upgrade the standard library and the interpreter separately sometimes, rather than in one big chunk."
Ok, you have a point here. This badness can do some harm.
"I like the idea of checked exceptions in some situations, but forcing every method to deal with (catching or throwing) all exceptions that its child calls or may call can be tedious. I'd rather be able to ignore an exception and let it propagate upwards. Sometimes, I'd rather not worry about exceptions at all."
My man, you know shit.
public Shit someMfkinMeth (Shit shitbooooy) throws throwable {
}
It's always interesting how people argue about this without reference to what we know about how things are in our world.
It is the case that states of afairs are represented by sets of symbols. It is the case that propositions, or truths are the result of states of afairs. Now, an inference is a relation between two propositions, and its nature is only understood from the nature of the two propositions itself. All rules of inference are irrelevant, because the nature of the inferences that are true is determined a-priori, for example the fact that my chair is on the floor is not derived from an understanding of the laws of gravity, but rather from the arrangement and nature of my chair and the floor. No elementary proposition can be deduced from another one. There is no way of inferencing from the existance of one sitation to the existance of another different situation. There is no causality. There simply is, an is.
Now, freedom of will is derived from the impossibility of knowing actions that lie in the future, while Strong AI is the belief that processes that manipulate physical devices of some sort are capable of actually becoming intelligent; in short that inferences between situations can be made. This may be true, but the only possible class of these inferences is the class of inferences that are true a-priori; thus the only inferences that an AI can make will be those that it can be predicted (possibly it will be very difficult) to make. Thus a machine of any type can appear to be intelligent, but it cannot have free will, and therefore it will be incapable of changing the world in which it acts; beyond the changes that could be directly infered from its introduction into that world. For example, a asteroid striking London would not change global politics, because the asteroid was always going to strike London, however the fact that people built London where it is and then it has become significant to other people is the thing that makes it's (putative) destruction significant
There is a difference, and if you wrap your silly little mid western mind around a half decent book like Tractatus Logico-Philosophicus you will figure it out!
My point?
Don't worry, this ain't going to happen, because it simply can't happen in this reality.
I would like to see a system where you were able to have two mail boxes - one that was public and caught all unsolicited mail. This could have a list of authorised users, and could forward mail from them to you. Non-authorised users could buy a stamp as suggested here, and then the public mail box could forward it on.
This would mean that people could pay to contact you, for example to change address or get in touch. Your friends and close collegues could use email as in the good old days.
Sure some of the demos are excellent. The problem is these implementations require just as much (at least) work as engineered solutions; it's just a different point of effort! Basically you have to discover how to make the self organising system organise itself to solve your problem, otherwise - it'll just sit and look at you. The trouble with this is that you don't know if it will work, until it works - a big risk for a commercial development project.
Some software is literally not worth marketing, it can be reverse engineered, can't support a business model because it is only useful for 100 people world wide or is background value - say worth 50cents a user a year. This is not to say that it is worthless, just not worth a marketing budget and a business. Giving away software like this is economically good, it saves many people time. Allowing those people to use it in commercial products is ecomonically better; rapid development of high quality stuff is enabled.
What we need to do is work out a system for moving some of the money back to the original contributers so that things are fair, and people are encouraged to do this. Think : you spend three wet weekends doing something, someone takes it and makes $millions, you never see a cent: fair? No! Should you get the $millions? NO! But you should see something for your skills.