I didn't know that. I have worked with Universities getting DARPA funding and my understanding was that they tended to spend the money as if it were a grant, but that the actually legalities of the matter were that it was a contract (just one with nebulous deliverables). Are you really sure they give grants?
No, Theo saying there were no strings attached in no way makes points 2 and 3 invalid. The government has its procedures and that's that. You simply can't take DARPA's money and not expect to write progress reports, explain what you're going to do with the money, etc., etc. In fact, I'd be surprised if Theo even considered writing progress reports, submitting accounts for the money received, etc. as "strings attached" --- that's just management and accountability.
In particular, DARPA issues contracts; DARPA does not issue grants. They are buying something from you and there are lots of rules about what the government must ask for when it buys stuff from people. You might believe there are no strings attached, but you'd be wrong.
As an aside, the same people who are outraged that Theo's money went away would probably be just as outraged if they found out that the government was spending money on something they didn't like without these controls. $700 toilet seats, anyone?
I very much doubt that this was pulled because of Theo's comments. I'd be absolutely shocked if I found out that one of UPenn's contract monitors actually read the Globe & Mail, stumbled on Theo's comments, and bothered to trace back to find out that he was funded by DARPA.
in 100 years, LISPers will be finally agree with different shapes of brackets. In fact they will accept that something like (defbracket {} metafunctor...) will let possible something like this (abc {x y z})
Errr.... Actually, you can already do this with Common Lisp today. If you have one handy, crack open a copy of Common Lisp: The Language, Second Edition and have a look at the discussion of reader macros. There are some examples of xappings and other things involving the Connection Machine, where new bracketings are used.
With all due respect, this doesn't seem like very interesting self awareness.
How about self awareness of programs, so that if you composed two programs, they could detect resource conflicts, and automatically add the appropriate semaphores (or whatever other locking you needed)?
Detecting other devices and talking to them is easy, compared to the problems of inferring how they might interact, and managing those interactions.
As far as the highway examples are concerned, we could probably do most of these today. The problems have to do with the open systems, the mixture of people and automatic cars on the road, and the unpleasant hardware issues like sensing the actual location of the road....
The PDA example we could do today, too, except for the fact that we don't have the social infrastructure (micropayments, etc.).
Freeman Dyson's book Weapons and Hope might be a good place to look. It's a little dated now, very much a cold war book, but it discusses Dyson's scientific role in the bombing of Germany, and the cold war.
Another possibility would be to discuss some of the more appalling medical studies. Consider the Japanese experiments on POWs that, uncomfortably enough, form the basis for modern treatment of pressure sickness. One might also discuss the Tuskegee experiments.
I'm not sure about the above two though, because they seem like slam-dunk examples of scientists doing Bad Things. One could do those, but it might be more interesting to cover a case that's more on the borderline, to make students really grapple with the issues instead of having a good feeling about recognizing an obvious case.
I like this a lot better, because it also makes it clear that Jews were not terrorists. And, for that matter, emphasizes that Arabs are not all terrorists, something we must all be careful to remember in these trying times.
Actually, that's not true. Readers do perform pattern recognition. When a fluent reader reads a written page, a lot of what's going on is pattern recognition. There isn't actually that much reading words out letter-by-letter.
For many things, words provide a much better discriminator, because you can more precisely capture meaning, and because there may be no good image for the process in question.
I have small children. Shiny disks just end up with little fingerprints all over them and in the wrong boxes. Anyway, I have too many to keep in easy reach. So I ripped all my CDs. I didn't pirate any. If I had a good DVD-ripper, I'd rip all the kids' videos (those are the ones we watch repeatedly). And ripping them would let me tear off the blankety-blank un-skippable commercials they put on to brainwash my kids.
Argh! Honestly, I'm getting more and more down with the guys who say "you should just have used lisp." With lisp you could have done that without popping it out and recompiling the JAR file. You could have hot-patched the executing code. and any half-decent modern common-lisp would have had advice or fwrappers to let you grab data without messing with the core functions.
Sure is great to see that one can plunder the past, introduce a new buzzword for the old idea, and supercharge one's career with a new programming paradigm.
Not just closures and macros --- how could you forget ADVICE, the construct that seems to be at the (unacknowledged?) core of this whole Aspect Oriented Paradigm ("My project's a paradigm, too!" - Dilbert).
yes, the military knows about security but they're now on this "commercial-off-the-shelf" kick, and it's not at all clear they fully understand the sacrifices they are making in terms of security and reliability when they do this. Seems odd. If you're going to buy a zillion cards and chips, wouldn't it still be relatively cheap to use a standard that's better secured than 802.11?
I was wondering --- am I the only one who's worried about what will happen when I "just type./install.sh"?
The scripts they write will, IIUC, uninstall packages they believe to conflict with theirs. But what if I have a specially-configured package that I want to keep? For example, I went out of my way, when installing the RealAudio player, to not let its installer take over all of the MIME types it wanted. I'm not sure I want to have them "fix" that.
Anyone tried an install that's not "for the rest of us"?
I agree. I think that this kind of thing shows the pernicious effects of the new "nothing matters but shareholder value" philosophy of corporate governance. It's time to bring some notion of corporate responsibility back to business ethics.
Concern for shareholders was a welcome corrective to managerial lack of accountability, but it's gone too far, and now it's time for the pendulum to start swinging the other way, towards concern for community and stakeholders other than just shareholders.
Meanwhile, I'm glad I work for a company that isn't publicly traded!
Actually, the advertised purpose of the ontology projects was to permit interchange of information between different intelligent systems. Ideally, the definitional information contained in the ontology was to permit different reasoning systems to understand each other's representations and communications. I would say the jury is still out on the utility of this approach.
There have long been systems that attempted to describe objects and their relationships (all the way back to Minsky's frames in the late 60s or early 70s. The ontologies were supposed to have some semantic groundwork that would let them be interchangeable and allow translation between different ontologies.
Similarly, on the web, ontologies for web page contents were intended to permit programs to grok the meaning of those web pages.
Unfortunately, a lot of effort seems to have gone into syntax and semantics of the ontologies themselves. This leaves some holes in:
The process of translating between ontologies
Figuring out where all that semantically marked-up content is going to come from.
For someone who's a little interested in Extreme Programming, would this be a reasonable first book?
If it were possible, I'd love to make my first book about Extreme Programming be one that took its limitations into account, discussed pros and cons, etc.
--
Somebody put my cheese under their copy of Extreme Programming
Finally, people like to criticise Israel and read off a memorized list of some dozen-odd UN security resolutions against Israel. Firstly, nearly all of these have provisions that the Palestinians too must adhere too, which they aren't. So it's BOTH Israel and Palestine in violation. Secondly, the entire Arab League is unilaterally unified against Israel (it was created strictly in opposition to the creation of Israel, but now it seems to be a valid entity), and have a significant block of power at the UN. To a lesser extent, the OIC (Organization of Islamic Countries) is nearly organized against Israel, and this is a block of about 50 votes in the 200-odd votes at the UN. So when people complain about Israel not following UN resolutions, it's important to keep in mind that a good block of the UN is specifically biased against Israel itself.
It's also worth taking into consideration, when reeling off those UN resolutions, the bias of the UN against Israel. For example, did you know that Israel's the only member state that can never be a member of the Security Council?
A UN General Assembly resolution against Israel is like an old Soviet Politburo resolution against the United States.
And if you think getting out of the Occupied Territories would solve the problem of conflict with the Palestinians, you might want to ask yourself why getting out of Lebanon didn't solve the problem of conflict with Hezbollah.
Sadly, the Real Media streams coming off the MIT site seem to be totally unreliable. Even across a DSL line, I'm not able to get through 10 minutes without the connection (and realplayer) going belly-up. Groan. So near, yet so far.
I thought @Large was quite interesting. It does a good job of debunking the "genius geek hero," too, showing the cracker as pretty pathetic, but also showing how much damage he could achieve. Good antidote to too much William Gibson (I love the stuff, but it's no more real than Raymond Chandler's about real private eyes).
One argument against having a homogeneous (uniform) environment is the argument from the security standpoint.
The argument here is pretty much the same as the argument against monoculture agriculture (which is also very convenient to administer): if you get a disease (in this case a virus or worm) that devastates a single species/configuration, you're dead.
Now, of course, one has to do a tradeoff --- maybe your stuff isn't so critical, and it makes better cost/benefit to just take the chance that the next Code Red will toast you, and you'll have to do an entire rebuild. But maybe your application is critical to safety, to your business, or whatever. In that case, you might want to seriously think about having a heterogeneous enterprise so that when your machines in configuration X are toasted, those in configuration Y can carry on until a fix is found.
Given that there are people out there who seem reasonably serious about attacking the U.S., maybe the DOI should think about this.
Given that there might be people interested in looting American Indian trust funds might be another reason....
Is this DSL or cable? As I said, my experience is only with DSL. I've heard of some folks who've had some glitches setting up cable, but not so many other annoyances (relatively frequent downtimes, etc.) as DSL.
I can't speak for cable, but where DSL is concerned, if I didn't need it to work from home, I'd be using dial-up.
I can't say too much about what a nightmare from hell my local telco (Qwest) has made DSL Until they make this significantly easier, change their dreadful customer service policies, make it work reliably and make it predictable (don't just spout vaguenesses about how it might work: figure out how to check the lines and make a commitment to deliver), the masses will not be ready for this.
Right now DSL is like driving one of the first cars: unless you're a mechanic, don't try.
Slashdot Mirror
I didn't know that. I have worked with Universities getting DARPA funding and my understanding was that they tended to spend the money as if it were a grant, but that the actually legalities of the matter were that it was a contract (just one with nebulous deliverables). Are you really sure they give grants?
No, Theo saying there were no strings attached in no way makes points 2 and 3 invalid. The government has its procedures and that's that. You simply can't take DARPA's money and not expect to write progress reports, explain what you're going to do with the money, etc., etc. In fact, I'd be surprised if Theo even considered writing progress reports, submitting accounts for the money received, etc. as "strings attached" --- that's just management and accountability.
In particular, DARPA issues contracts; DARPA does not issue grants. They are buying something from you and there are lots of rules about what the government must ask for when it buys stuff from people. You might believe there are no strings attached, but you'd be wrong.
As an aside, the same people who are outraged that Theo's money went away would probably be just as outraged if they found out that the government was spending money on something they didn't like without these controls. $700 toilet seats, anyone?
I very much doubt that this was pulled because of Theo's comments. I'd be absolutely shocked if I found out that one of UPenn's contract monitors actually read the Globe & Mail, stumbled on Theo's comments, and bothered to trace back to find out that he was funded by DARPA.
Errr.... Actually, you can already do this with Common Lisp today. If you have one handy, crack open a copy of Common Lisp: The Language, Second Edition and have a look at the discussion of reader macros. There are some examples of xappings and other things involving the Connection Machine, where new bracketings are used.
With all due respect, this doesn't seem like very interesting self awareness.
How about self awareness of programs, so that if you composed two programs, they could detect resource conflicts, and automatically add the appropriate semaphores (or whatever other locking you needed)?
Detecting other devices and talking to them is easy, compared to the problems of inferring how they might interact, and managing those interactions.
As far as the highway examples are concerned, we could probably do most of these today. The problems have to do with the open systems, the mixture of people and automatic cars on the road, and the unpleasant hardware issues like sensing the actual location of the road....
The PDA example we could do today, too, except for the fact that we don't have the social infrastructure (micropayments, etc.).
With a bridge to 802.11b it gets really expensive: over $300 :-
Freeman Dyson's book Weapons and Hope might be a good place to look. It's a little dated now, very much a cold war book, but it discusses Dyson's scientific role in the bombing of Germany, and the cold war.
Another possibility would be to discuss some of the more appalling medical studies. Consider the Japanese experiments on POWs that, uncomfortably enough, form the basis for modern treatment of pressure sickness. One might also discuss the Tuskegee experiments.
I'm not sure about the above two though, because they seem like slam-dunk examples of scientists doing Bad Things. One could do those, but it might be more interesting to cover a case that's more on the borderline, to make students really grapple with the issues instead of having a good feeling about recognizing an obvious case.
I like this a lot better, because it also makes it clear that Jews were not terrorists. And, for that matter, emphasizes that Arabs are not all terrorists, something we must all be careful to remember in these trying times.
Actually, that's not true. Readers do perform pattern recognition. When a fluent reader reads a written page, a lot of what's going on is pattern recognition. There isn't actually that much reading words out letter-by-letter.
For many things, words provide a much better discriminator, because you can more precisely capture meaning, and because there may be no good image for the process in question.
I have small children. Shiny disks just end up with little fingerprints all over them and in the wrong boxes. Anyway, I have too many to keep in easy reach. So I ripped all my CDs. I didn't pirate any. If I had a good DVD-ripper, I'd rip all the kids' videos (those are the ones we watch repeatedly). And ripping them would let me tear off the blankety-blank un-skippable commercials they put on to brainwash my kids.
wireless-p?
Argh! Honestly, I'm getting more and more down with the guys who say "you should just have used lisp." With lisp you could have done that without popping it out and recompiling the JAR file. You could have hot-patched the executing code. and any half-decent modern common-lisp would have had advice or fwrappers to let you grab data without messing with the core functions.
Sure is great to see that one can plunder the past, introduce a new buzzword for the old idea, and supercharge one's career with a new programming paradigm.
Not just closures and macros --- how could you forget ADVICE, the construct that seems to be at the (unacknowledged?) core of this whole Aspect Oriented Paradigm ("My project's a paradigm, too!" - Dilbert).
yes, the military knows about security but they're now on this "commercial-off-the-shelf" kick, and it's not at all clear they fully understand the sacrifices they are making in terms of security and reliability when they do this. Seems odd. If you're going to buy a zillion cards and chips, wouldn't it still be relatively cheap to use a standard that's better secured than 802.11?
I was wondering --- am I the only one who's worried about what will happen when I "just type ./install.sh"?
The scripts they write will, IIUC, uninstall packages they believe to conflict with theirs. But what if I have a specially-configured package that I want to keep? For example, I went out of my way, when installing the RealAudio player, to not let its installer take over all of the MIME types it wanted. I'm not sure I want to have them "fix" that.
Anyone tried an install that's not "for the rest of us"?
Actually, if you are a military contractor, it is illegal for you to have poor security for your buildings, and furthermore, it ought to be illegal!
I agree. I think that this kind of thing shows the pernicious effects of the new "nothing matters but shareholder value" philosophy of corporate governance. It's time to bring some notion of corporate responsibility back to business ethics.
Concern for shareholders was a welcome corrective to managerial lack of accountability, but it's gone too far, and now it's time for the pendulum to start swinging the other way, towards concern for community and stakeholders other than just shareholders.
Meanwhile, I'm glad I work for a company that isn't publicly traded!
Actually, the advertised purpose of the ontology projects was to permit interchange of information between different intelligent systems. Ideally, the definitional information contained in the ontology was to permit different reasoning systems to understand each other's representations and communications. I would say the jury is still out on the utility of this approach.
There have long been systems that attempted to describe objects and their relationships (all the way back to Minsky's frames in the late 60s or early 70s. The ontologies were supposed to have some semantic groundwork that would let them be interchangeable and allow translation between different ontologies.
Similarly, on the web, ontologies for web page contents were intended to permit programs to grok the meaning of those web pages.
Unfortunately, a lot of effort seems to have gone into syntax and semantics of the ontologies themselves. This leaves some holes in:
For someone who's a little interested in Extreme Programming, would this be a reasonable first book?
If it were possible, I'd love to make my first book about Extreme Programming be one that took its limitations into account, discussed pros and cons, etc.
--Somebody put my cheese under their copy of Extreme Programming
Are we sure that this is true? I would have thought that something like Trustix or Immunix might be certified. Are they not?
It's also worth taking into consideration, when reeling off those UN resolutions, the bias of the UN against Israel. For example, did you know that Israel's the only member state that can never be a member of the Security Council?
A UN General Assembly resolution against Israel is like an old Soviet Politburo resolution against the United States.
And if you think getting out of the Occupied Territories would solve the problem of conflict with the Palestinians, you might want to ask yourself why getting out of Lebanon didn't solve the problem of conflict with Hezbollah.
Sadly, the Real Media streams coming off the MIT site seem to be totally unreliable. Even across a DSL line, I'm not able to get through 10 minutes without the connection (and realplayer) going belly-up. Groan. So near, yet so far.
I thought @Large was quite interesting. It does a good job of debunking the "genius geek hero," too, showing the cracker as pretty pathetic, but also showing how much damage he could achieve. Good antidote to too much William Gibson (I love the stuff, but it's no more real than Raymond Chandler's about real private eyes).
One argument against having a homogeneous (uniform) environment is the argument from the security standpoint.
The argument here is pretty much the same as the argument against monoculture agriculture (which is also very convenient to administer): if you get a disease (in this case a virus or worm) that devastates a single species/configuration, you're dead.
Now, of course, one has to do a tradeoff --- maybe your stuff isn't so critical, and it makes better cost/benefit to just take the chance that the next Code Red will toast you, and you'll have to do an entire rebuild. But maybe your application is critical to safety, to your business, or whatever. In that case, you might want to seriously think about having a heterogeneous enterprise so that when your machines in configuration X are toasted, those in configuration Y can carry on until a fix is found.
Given that there are people out there who seem reasonably serious about attacking the U.S., maybe the DOI should think about this.
Given that there might be people interested in looting American Indian trust funds might be another reason....
Is this DSL or cable? As I said, my experience is only with DSL. I've heard of some folks who've had some glitches setting up cable, but not so many other annoyances (relatively frequent downtimes, etc.) as DSL.
I can't speak for cable, but where DSL is concerned, if I didn't need it to work from home, I'd be using dial-up.
I can't say too much about what a nightmare from hell my local telco (Qwest) has made DSL Until they make this significantly easier, change their dreadful customer service policies, make it work reliably and make it predictable (don't just spout vaguenesses about how it might work: figure out how to check the lines and make a commitment to deliver), the masses will not be ready for this.
Right now DSL is like driving one of the first cars: unless you're a mechanic, don't try.