While MS has been banged around enough on slashdot, for the most part MS has really been more of a positive influence than otherwise for most users. I'd much prefer more diversity in computer offerings (and in computer education - in my university "computer literacy" means "microsoft literacy"), I'd like more open drivers, more standardization - ah, you all know the problems. But for most users MS has been a Good Thing (with some qualifications).
But there is another thought that sometimes wanders through my brain - is it not possible that MS has been pricing things low in large part because of the steering provided by Bill Gates? I've read a few things that have suggested that Gates does indeed push the company in the direction of keeping prices down.
In which case, when Gates leaves the company, or if (by some marvelous melange of machinations) someone managed to buy a controlling interest. Might not new management decide that the prices are too low? That the DRM is not strict enough? That their XML should become completely proprietary again? With their monopoly, many businesses (and lots of consumers) could easily find themselves paying monthly subscription fees that might make their current fee structure look paltry. And, as has been said here so many times, many businesses would have little choice but to pay extortionate fees - and MS would be supported by the hordes of MS only developers who'd probably figure that they could hike their fees proportionately.
Clearly Google could do the same - but Google is operating in a very different market, with lots more competition (though with the product discussed here, they're clearly taking on MS in a key business area).
A messy question indeed. No declaration of war was ever passed, but then there is some dispute about what that actually means. There also seems to be some interesting manoeuvering involving the war status - sometimes the US government seems to find it convenient to think the US is at war, then at other times it is more convenient to say the country is not at war. A frequent example is when the question of "detainees" vs "prisoners of war" comes up.
Of course, any smart occupying force will ensure that the government of the occupied region signs all the right agreements.
Eventually it all comes down to an ethical (rather than a legal) question, and governments (pretty much everywhere) don't seem to have much in the way of ethics. But then can governments (as opposed to individuals) really have ethics? But then too, many individuals in government don't.
It seems to me that criminal prosecutions of SCO's legal team would (even if something that the law allows) set a precedent that would make it easier for large/rich corporations to seriously threaten smaller corporations or individuals who want to challenge them in court. Perhaps there could be Iis?) some kind of middle ground that would allow any party to a lawsuit to file a complaint that would result in fines, censure or some other less severe punishment for the legal team.
Similarly, putting SCO officers in jail on the basis of their legal case would have the same effect. But if it could be proved that they brought the case to manipulate their stock prices, or if they committed perjury at any point along the way, jail time would be more than appropriate.
"In advertising there is no such thing as a lie, there is only expedient exaggeration."
Roger O Thornhill (played by Cary Grant) in "North by Northwest".
Much as I thought that getting in to Iraq was at best stupid, and much as I think the US really should not be there, simply getting out seems less than responsible. We replaced an insane, cruel, arbitrary dictator who nonetheless kept order (but, of course, for how long) with chaos, foreign occupation and what might seems to be turning out to be an insane, cruel , arbitrary and very disorderly civil war. If there is a way to do it, (which I'm far from sure is the case) we owe it to the Iraqis (and their neighbors and the international community in general) to leave the country with some kind of stability that is likely to last for a bit. Otherwise, we are not only not fixing anything, we have, like a spoiled child , deliberately broken things and are leaving the pieces for everyone else to pick up and try to fix.
It may not be feasible to find a way to manage this. It is almost certainly not going to be feasible without international cooperation which the US administration seems to continue to find distasteful. It doesn't help that they pissed off lots of other countries earlier in the process. It probably will involve a US president going with some serious humility to other countries, hat in hand, to ask for support (which almost certainly means it will not be this US president). But we owe it to the Iraqis to at least do our best to try.
We also owe it to ourselves. Leaving Iraq to fester in civil war will be a legacy that the US will find it hard to overcome. If the civil war involves other Middle Eastern countries, it is hard to see how the US will avoid being seen as the ultimate cause. (Certainly Saddam Hussein also shares a good bit of the blame, but then the US seems to have been at least partly responsible in putting him in power.)
It has been my experience that bureaucracies don't generally want to believe that their policies and procedures are, indeed, essentially software and could therefore be considered as such, with provisions for analysis and with their design including provisions for exceptional cases. It is tough to debug such things once they get deployed - in part because there are usually few ways to report "bugs" and often nobody to read such bug reports if they were generated. In the ideal, the policies would specifically allow for the odd cases and provide easy ways to handle the odd cases (by "throwing exceptions" up to a superior) but in practice that is often unworkable because the "superior" has other things to cope with. doesn't care, or has allowed their "superiority" to delude them into thinking their job does not include such things.
No matter how well planned and executed, all systems eventually do other things than they were designed to do (see "Systemantics" by John Gall).
(A bit of self promotion.) I took his idea and incorporated it into a genetic programming
system that manages to crash most browsers. It also finds HTML source that causes browsers to work for a looooonnnggg time to render a single page (in one case 19 hours for a page). The HTML is not particularly
legal, but then there is no guarantee that any web page you load into a browser will follow any particular
standard. Source
(Java) is available at sourceforge - unpack and look for subdirectory "html". (Warning: As this is an evolving program subject to random hackery to "enhance" things, it is commented sketchily, way underdocumented and far from pretty in most places.)
My first thought was exactly this - investigate machines made by that leftist radical country Venezuela, but don't even think about critizing those made by Diebold.
My second thought was way more cynical - could this be a pre-election manoeuvre designed to set the stage for investigating/recount/tossing out votes on suspect machines? Being sure, of course, to focus on those machines that voted Democrat. (OK, so the machines don't themselves vote. Or do they? Since we often can't see the the code or even verify the hardware, it is quite possible that the machines do themselves vote.)
I once did the "emerge --flag-that-tells-you-what-will-be-emerged somepackage" and nothing problematic showed up. I was feeling all set to do the emerge, but then had to handle something else for a few minutes (half hour or so), and returned and typed the "emerge somepackage" and went away to do something else in the meantime. Murphy's Law intervened, of course, and when the emerge started, libc had changed so there were several hundred (?) packages that needed to be upgraded. But that was only part of the problem. Libc got installed ok, I think, but then everything seemed to hang. I got back and tried to fix things, in the process only making things worse. Eventually I gave up and (since this was the second time gentoo had forced a complete install on me) installed debian. Happily all my personal stuff was in/home and survived ok.
Since running debian/ubuntu, I have had a couple minor problems (this machine is running a fresh install of eft) but nothing like that.
Couldn't you put a kernel module in place that would overwrite swap space from a process
with random data
when that process exits? Also, the same kind of thing for all of swap space when the
machine shuts down and reboots? (The shut down would probably impose a rather annoying
wait, but randomizing swap on reboot could probably be done in parallel with other startup
tasks.)
I'd be just as worried about temporary files if the/home filesystem were encrypted but
not/tmp, but I'd suspect that a userspace program that grabbed empty space in/tmp,
wrote random data on it, released it... repeat would probably handle it reasonably (at least
with high probability).
This is always the thing that I wonder about. So Windows pricing has been more or less reasonable up to now
(in large part because of the pre-installs from every OEM), but what if a new generation of folks were to
take over Microsoft, realize that their monopoly position makes it almost impossible for anyone to come out
with a viable competitor in any reasonable timeframe and then raise the price by some interesting factor
say to something like $500 for every OEM install and $2000 for corporate desktops.
Sure, there'd be some installs of Linspire and some people running MacOS (and maybe Apple would see this as
a good time to make MacOS install on lots of machines), but for hard core gamers and especially for corporate
IT departments, it would be next to impossible to switch quickly and they'd end up paying quite a premium.
Larger corporations with direct contracts with MS would not see the effect for a while, but smaller companies
would pretty much have to grin and bear it.
I suspect there'd be quite a bit of piracy, but a few lawsuits from Redmond would take care of that quickly
enough (I suspect there'd be a whole mini-ecology of lawyers getting rich on both sides). We already
know how well anti-monopoly laws work here.
I'm the original poster and I'm not sure I see anything wrong with doing this since the information is public. I found the article interesting (if a bit skimpy). A number of questions came to mind, but I only included the one on syndication since the article specifically said that the US press would not be monitored and that foreign press would. If things are syndicated, how do you tell the difference?
I didn't want to load the post with questions but I wondered in particular what natural language understanding techniques could provide that Google News (and other news aggregators) does not already provide - certainly the trending information available seems particularly interesting. I also wondered if the language understanding is really all that strong for other languages (including specifically Arabic since the lack of Arabic speakers has come up as newsworth a number of times) - especially in US universities. A deeper problem might be to distinguish between anti-american attitudes and anti-(american-president-and-administration)-attit udes. Finally, I found the mention that specific journalists attitudes might be tracked a bit troublesome - would this be used then to deny visas to journalists who consistently manifest anti-american(-president-and-administration)-attit udes.
Should the New Yorker not cover things that may be beyond the reach of the average reader?
Even if they were publishing the mathematical theory itself, they should be free to do so (though
it would probably not appeal to the average reader), but they're not doing that, they're publishing about
a controversy in the field - just as they might about any other field. Is physics somehow different than
(to take an example from one article I remember) considering the effectiveness of different kinds of therapy
on people who've experienced stressful events and who might then be subject to PTSD?
Writers and journalists should be encouraged to write about whatever interests them and their audience, even
if the people they're writing about don't always find it flattering or helpful.
As someone who frequently reads the New Yorker, I must say I've learned a lot from it over the years - and in
many areas that I'm not familiar with such readings have sometimes taught me something (perhaps only a little, but something),
sometimes aroused my curiousity, and sometimes introduced me to whole new ideas that I might not have otherwise
run into. I say "More power to 'em".
First, the "New Yorker" is not a paper, its a magazine.
Second, while the "New Yorker" has, on occasion, really made news (particularly noted are
Seymore Hirsh's articles on Iraq and related policies), for the most part it does do reporting.
I believe that the skepticism about string theory has been around in physics circles for a while, and
for the magazine to report on this hardly makes news - it just tries to make news about physics (and
mathematics) accessible to the general reader. Certainly the story about the Poincare conjector was
not created out of thin air - but since it was covered in a wide circulation magazine, it came to the
attention of many more people than would otherwise have heard of it. But does that count as "making
news"? I doubt it.
While the New Yorker article was not particularly favorable to Dr. Yau, it didn't seem to me that it could be called defamation. Indeed, to the extent that
it says negative things about him, they seem to be coming from his peers in mathematics - and not from the writer of the article. Is that a sufficient
defense against a legal claim of defamation? I guess that is for the courts to decide.
More importantly, by suing for defamation, Dr. Yau appears to be manifesting exactly the kind of behavior that he was described as having in the article. One mathematician is
quoted as saying "Yau wants to be the king of geometry. He believes that everything should issue from him, that he should have oversight. He doesn't like people encroaching on his territory.". Another says : "This is a guy who did magnificent things... He won every prize to be won. I find it a little mean of him to seem to be trying to get a share of this as well."
SML or OCaml are great lenguages, but if you're going to learn a functional language, Haskell is a great place to start. First because the syntax is very clean (I never quite liked the "let rec" bit) but also because in both SML and OCaml it is too easy to slip back into imperative styles. Haskell makes that substantially more difficult, which means to use it well you really, really have to get the whole functional programming idea.
I don't think (as a college professor) that written exams do much to really reflect what a student either knows, or is capable of. There are exceptions - it is possible to write an exam that does a decent job of measuring a students real abilities, but it is very, very difficult, and it is equally difficult to score.
An oral exam, lasting a half hour to an hour, will usually give much more information, and would provide the students with valuable feedback. It also works a bit like those computerized exams where you only get enough questions to place you (as accurately as possible) on a scoring scale.
I'd love to give oral exams to students, but there are several major problems - first, with any number of students, it can be very difficult to schedule. Even with a 30 student class, orals would take 15-30 hours to manage and this at the end of the term/semester where you have other exams, grading and the like. Secondly, and perhaps far more importantly, orals (as well as the kind of flexible written test mentioned above) are almost completely subjective and thus probably unacceptable to most students. The trend these days is toward specific "goals and objectives" for a course with measurable and definable "outcomes and assessments". Something like "can use inheritance in an Object Oriented Language" is becoming unacceptably vague. Something more like "Given a Java class, the student can define a subclass with additional methods and fields." is likely to be more acceptable, but with this trend we're now seeing people demanding even more precise specifications ("Given the java class ArrayList, the student can build a subclass with an iterator that generates the elements in random order"). This makes writing the test easier, but also means that teaching to the test is not only inevitable but demanded.
As someone said above, try Squeak. The debugger isn't structured quite like the usual (gdb type) debuggers, but is wonderfully powerful and you can even change the code while in the debugger and ask it to try to continue from where it left off.
Random searches are probably the best way to do things - but they should indeed be random. Toss a
twenty sided die for each passenger perhaps, or compute a hash of their ticket number XORed with the
time and use that as a basis for deciding.
If profiling of any sort is used, then suppose the bad guys have a group of twenty people who are
potential hijackers or bombers or the like. Then let each of them travel several times before their planned
attack and see who is searched. The rest probably do not match the profile and can be used for the
attack. Ideally, of course, the profiling will catch all of them, but given a big enough pool of
bad guys this becomes less and less likely.
The random method is fair, effective (as there is no predictable way for the bad guys to avoid it) and probably
as effective as any other method you might imagine.
In the Congo, there are a number of tribal languages (a couple of hundred, if I remember correctly) and several major trade languages
that are common across large regions (I was in the Peace Corps there a ways back and my electricity bill came in seven languages).
But Mobuto (President at the time) spoke Lingala and was pushing it hard as the primary official language. The people in the eastern part of the country (where Kiswahili was the lingua franca) resented it more than a bit, and especially resented the administrators who would come to the area and who spoke no
Kiswahili at all. Of course, this is linked in with tribalism as well as resentment of Mobutu (who was not a nice person).
As a result, the common language that really unified the country was French (which most educated people spoke quite well).
Slashdot Mirror
While MS has been banged around enough on slashdot, for the most part MS has really been more of a positive influence than otherwise for most users. I'd much prefer more diversity in computer offerings (and in computer education - in my university "computer literacy" means "microsoft literacy"), I'd like more open drivers, more standardization - ah, you all know the problems. But for most users MS has been a Good Thing (with some qualifications).
But there is another thought that sometimes wanders through my brain - is it not possible that MS has been pricing things low in large part because of the steering provided by Bill Gates? I've read a few things that have suggested that Gates does indeed push the company in the direction of keeping prices down. In which case, when Gates leaves the company, or if (by some marvelous melange of machinations) someone managed to buy a controlling interest. Might not new management decide that the prices are too low? That the DRM is not strict enough? That their XML should become completely proprietary again? With their monopoly, many businesses (and lots of consumers) could easily find themselves paying monthly subscription fees that might make their current fee structure look paltry. And, as has been said here so many times, many businesses would have little choice but to pay extortionate fees - and MS would be supported by the hordes of MS only developers who'd probably figure that they could hike their fees proportionately.
Clearly Google could do the same - but Google is operating in a very different market, with lots more competition (though with the product discussed here, they're clearly taking on MS in a key business area).
Is the US legally at war with Iraq or not?
A messy question indeed. No declaration of war was ever passed, but then there is some dispute about what that actually means. There also seems to be some interesting manoeuvering involving the war status - sometimes the US government seems to find it convenient to think the US is at war, then at other times it is more convenient to say the country is not at war. A frequent example is when the question of "detainees" vs "prisoners of war" comes up.
Of course, any smart occupying force will ensure that the government of the occupied region signs all the right agreements.
Eventually it all comes down to an ethical (rather than a legal) question, and governments (pretty much everywhere) don't seem to have much in the way of ethics. But then can governments (as opposed to individuals) really have ethics? But then too, many individuals in government don't.
It seems to me that criminal prosecutions of SCO's legal team would (even if something that the law allows) set a precedent that would make it easier for large/rich corporations to seriously threaten smaller corporations or individuals who want to challenge them in court. Perhaps there could be Iis?) some kind of middle ground that would allow any party to a lawsuit to file a complaint that would result in fines, censure or some other less severe punishment for the legal team.
Similarly, putting SCO officers in jail on the basis of their legal case would have the same effect. But if it could be proved that they brought the case to manipulate their stock prices, or if they committed perjury at any point along the way, jail time would be more than appropriate.
"In advertising there is no such thing as a lie, there is only expedient exaggeration."
Roger O Thornhill (played by Cary Grant) in "North by Northwest".
Much as I thought that getting in to Iraq was at best stupid, and much as I think the US really should not be there, simply getting out seems less than responsible. We replaced an insane, cruel, arbitrary dictator who nonetheless kept order (but, of course, for how long) with chaos, foreign occupation and what might seems to be turning out to be an insane, cruel , arbitrary and very disorderly civil war. If there is a way to do it, (which I'm far from sure is the case) we owe it to the Iraqis (and their neighbors and the international community in general) to leave the country with some kind of stability that is likely to last for a bit. Otherwise, we are not only not fixing anything, we have, like a spoiled child , deliberately broken things and are leaving the pieces for everyone else to pick up and try to fix.
It may not be feasible to find a way to manage this. It is almost certainly not going to be feasible without international cooperation which the US administration seems to continue to find distasteful. It doesn't help that they pissed off lots of other countries earlier in the process. It probably will involve a US president going with some serious humility to other countries, hat in hand, to ask for support (which almost certainly means it will not be this US president). But we owe it to the Iraqis to at least do our best to try.
We also owe it to ourselves. Leaving Iraq to fester in civil war will be a legacy that the US will find it hard to overcome. If the civil war involves other Middle Eastern countries, it is hard to see how the US will avoid being seen as the ultimate cause. (Certainly Saddam Hussein also shares a good bit of the blame, but then the US seems to have been at least partly responsible in putting him in power.)
It has been my experience that bureaucracies don't generally want to believe that their policies and procedures are, indeed, essentially software and could therefore be considered as such, with provisions for analysis and with their design including provisions for exceptional cases. It is tough to debug such things once they get deployed - in part because there are usually few ways to report "bugs" and often nobody to read such bug reports if they were generated. In the ideal, the policies would specifically allow for the odd cases and provide easy ways to handle the odd cases (by "throwing exceptions" up to a superior) but in practice that is often unworkable because the "superior" has other things to cope with. doesn't care, or has allowed their "superiority" to delude them into thinking their job does not include such things.
No matter how well planned and executed, all systems eventually do other things than they were designed to do (see "Systemantics" by John Gall).
To quote a very wise man :
"Guns don't kill people, physics kills people." - Dr Dick Solomon
Or in this case :
"Capacitors don't kill people,... "
Just crashing browsers is easy enough. Even just with HTML. Remember this story?
(A bit of self promotion.) I took his idea and incorporated it into a genetic programming system that manages to crash most browsers. It also finds HTML source that causes browsers to work for a looooonnnggg time to render a single page (in one case 19 hours for a page). The HTML is not particularly legal, but then there is no guarantee that any web page you load into a browser will follow any particular standard. Source (Java) is available at sourceforge - unpack and look for subdirectory "html". (Warning: As this is an evolving program subject to random hackery to "enhance" things, it is commented sketchily, way underdocumented and far from pretty in most places.)
Nice!
That may be something worth submitting to wikiquote.
My first thought was exactly this - investigate machines made by that leftist radical country Venezuela, but don't even think about critizing those made by Diebold.
My second thought was way more cynical - could this be a pre-election manoeuvre designed to set the stage for investigating/recount/tossing out votes on suspect machines? Being sure, of course, to focus on those machines that voted Democrat. (OK, so the machines don't themselves vote. Or do they? Since we often can't see the the code or even verify the hardware, it is quite possible that the machines do themselves vote.)
I once did the "emerge --flag-that-tells-you-what-will-be-emerged somepackage" and nothing problematic showed up. I was feeling all set to do the emerge, but then had to handle something else for a few minutes (half hour or so), and returned and typed the "emerge somepackage" and went away to do something else in the meantime. Murphy's Law intervened, of course, and when the emerge started, libc had changed so there were several hundred (?) packages that needed to be upgraded. But that was only part of the problem. Libc got installed ok, I think, but then everything seemed to hang. I got back and tried to fix things, in the process only making things worse. Eventually I gave up and (since this was the second time gentoo had forced a complete install on me) installed debian. Happily all my personal stuff was in /home and survived ok.
Since running debian/ubuntu, I have had a couple minor problems (this machine is running a fresh install of eft) but nothing like that.
I'd be just as worried about temporary files if the /home filesystem were encrypted but
not /tmp, but I'd suspect that a userspace program that grabbed empty space in /tmp,
wrote random data on it, released it... repeat would probably handle it reasonably (at least
with high probability).
This is always the thing that I wonder about. So Windows pricing has been more or less reasonable up to now (in large part because of the pre-installs from every OEM), but what if a new generation of folks were to take over Microsoft, realize that their monopoly position makes it almost impossible for anyone to come out with a viable competitor in any reasonable timeframe and then raise the price by some interesting factor say to something like $500 for every OEM install and $2000 for corporate desktops.
Sure, there'd be some installs of Linspire and some people running MacOS (and maybe Apple would see this as a good time to make MacOS install on lots of machines), but for hard core gamers and especially for corporate IT departments, it would be next to impossible to switch quickly and they'd end up paying quite a premium. Larger corporations with direct contracts with MS would not see the effect for a while, but smaller companies would pretty much have to grin and bear it.
I suspect there'd be quite a bit of piracy, but a few lawsuits from Redmond would take care of that quickly enough (I suspect there'd be a whole mini-ecology of lawyers getting rich on both sides). We already know how well anti-monopoly laws work here.
More like the heart of propaganda - and it is used by all sides. In particular, the parent uses exactly this technique to redefine liberalistm.
I think you mean CALCULUS by Michael Spivak, and I agree, it is a great calculus book.
I didn't want to load the post with questions but I wondered in particular what natural language understanding techniques could provide that Google News (and other news aggregators) does not already provide - certainly the trending information available seems particularly interesting. I also wondered if the language understanding is really all that strong for other languages (including specifically Arabic since the lack of Arabic speakers has come up as newsworth a number of times) - especially in US universities. A deeper problem might be to distinguish between anti-american attitudes and anti-(american-president-and-administration)-attit udes. Finally, I found the mention that specific journalists attitudes might be tracked a bit troublesome - would this be used then to deny visas to journalists who consistently manifest anti-american(-president-and-administration)-attit udes.
Should the New Yorker not cover things that may be beyond the reach of the average reader?
Even if they were publishing the mathematical theory itself, they should be free to do so (though it would probably not appeal to the average reader), but they're not doing that, they're publishing about a controversy in the field - just as they might about any other field. Is physics somehow different than (to take an example from one article I remember) considering the effectiveness of different kinds of therapy on people who've experienced stressful events and who might then be subject to PTSD?
Writers and journalists should be encouraged to write about whatever interests them and their audience, even if the people they're writing about don't always find it flattering or helpful.
As someone who frequently reads the New Yorker, I must say I've learned a lot from it over the years - and in many areas that I'm not familiar with such readings have sometimes taught me something (perhaps only a little, but something), sometimes aroused my curiousity, and sometimes introduced me to whole new ideas that I might not have otherwise run into. I say "More power to 'em".
First, the "New Yorker" is not a paper, its a magazine.
Second, while the "New Yorker" has, on occasion, really made news (particularly noted are Seymore Hirsh's articles on Iraq and related policies), for the most part it does do reporting. I believe that the skepticism about string theory has been around in physics circles for a while, and for the magazine to report on this hardly makes news - it just tries to make news about physics (and mathematics) accessible to the general reader. Certainly the story about the Poincare conjector was not created out of thin air - but since it was covered in a wide circulation magazine, it came to the attention of many more people than would otherwise have heard of it. But does that count as "making news"? I doubt it.
While the New Yorker article was not particularly favorable to Dr. Yau, it didn't seem to me that it could be called defamation. Indeed, to the extent that it says negative things about him, they seem to be coming from his peers in mathematics - and not from the writer of the article. Is that a sufficient defense against a legal claim of defamation? I guess that is for the courts to decide.
More importantly, by suing for defamation, Dr. Yau appears to be manifesting exactly the kind of behavior that he was described as having in the article. One mathematician is quoted as saying "Yau wants to be the king of geometry. He believes that everything should issue from him, that he should have oversight. He doesn't like people encroaching on his territory.". Another says : "This is a guy who did magnificent things... He won every prize to be won. I find it a little mean of him to seem to be trying to get a share of this as well."
SML or OCaml are great lenguages, but if you're going to learn a functional language, Haskell is a great place to start. First because the syntax is very clean (I never quite liked the "let rec" bit) but also because in both SML and OCaml it is too easy to slip back into imperative styles. Haskell makes that substantially more difficult, which means to use it well you really, really have to get the whole functional programming idea.
An oral exam, lasting a half hour to an hour, will usually give much more information, and would provide the students with valuable feedback. It also works a bit like those computerized exams where you only get enough questions to place you (as accurately as possible) on a scoring scale.
I'd love to give oral exams to students, but there are several major problems - first, with any number of students, it can be very difficult to schedule. Even with a 30 student class, orals would take 15-30 hours to manage and this at the end of the term/semester where you have other exams, grading and the like. Secondly, and perhaps far more importantly, orals (as well as the kind of flexible written test mentioned above) are almost completely subjective and thus probably unacceptable to most students. The trend these days is toward specific "goals and objectives" for a course with measurable and definable "outcomes and assessments". Something like "can use inheritance in an Object Oriented Language" is becoming unacceptably vague. Something more like "Given a Java class, the student can define a subclass with additional methods and fields." is likely to be more acceptable, but with this trend we're now seeing people demanding even more precise specifications ("Given the java class ArrayList, the student can build a subclass with an iterator that generates the elements in random order"). This makes writing the test easier, but also means that teaching to the test is not only inevitable but demanded.
I'll bet they own lots of stock in an aspirin company somewhere.
As someone said above, try Squeak. The debugger isn't structured quite like the usual (gdb type) debuggers, but is wonderfully powerful and you can even change the code while in the debugger and ask it to try to continue from where it left off.
If profiling of any sort is used, then suppose the bad guys have a group of twenty people who are potential hijackers or bombers or the like. Then let each of them travel several times before their planned attack and see who is searched. The rest probably do not match the profile and can be used for the attack. Ideally, of course, the profiling will catch all of them, but given a big enough pool of bad guys this becomes less and less likely.
The random method is fair, effective (as there is no predictable way for the bad guys to avoid it) and probably as effective as any other method you might imagine.
Never underestimate the power of randomness!
In the Congo, there are a number of tribal languages (a couple of hundred, if I remember correctly) and several major trade languages that are common across large regions (I was in the Peace Corps there a ways back and my electricity bill came in seven languages). But Mobuto (President at the time) spoke Lingala and was pushing it hard as the primary official language. The people in the eastern part of the country (where Kiswahili was the lingua franca) resented it more than a bit, and especially resented the administrators who would come to the area and who spoke no Kiswahili at all. Of course, this is linked in with tribalism as well as resentment of Mobutu (who was not a nice person). As a result, the common language that really unified the country was French (which most educated people spoke quite well).