When a so called 'newbie' starts out he shouldn't (need to) read documentation telling him how to use vi to edit/etc/samba/smb.conf.
And it should be pointed out that even those of us whose preferred mode of configuration is something like that still depend on certain conventions, which can do a great deal to make the software more or less usable.
E.g., the configuration files should be located at a path in etc with "smb" or "samba" somewhere in the name, documentation should be deposited in the standard places, the configuration file should use a format that's also used by other programs instead of one you just made up, etc., etc.
Similarly gui configurators should be findable in some standard location, should use widgets users have seen before, should label options clearly, etc.
People focus so much on the newbie/sysadmin and gui/commandline dichotomy's that they forget that the problems faced by both are very similar.
I can't find an article reference to it on google news (might be too long ago), but a few months back there was an incident where a guy jumped off a Tri-Met bus and shot someone in broad daylight downtown.
And how many people died in this incident? And how many people pass through downtown Portland in the average day? If you were really worried about your chances of being killed while walking through downtown Portland, surely you'd be better off, say, working on your street-crossing technique.
Huh? I spent lots of time wandering around downtown Portland at all hours of the night and day (and lived at 16th and Yamhill for a year) in the early ninetees, and can't think of anything that would prompt this comment. It didn't look much different when I visted there last summer.
If you were going to feel weird about any area, OK, maybe just north of Burnside is a little seedy, but I never had any actual trouble there either.
> that doesn't necessarily mean that withholding such permits would have no effect on the availability of such guns to such people
Umm, actually, yes it does. If they aren't getting permits, and the only new restriction you impose is making permits harder to get, then the number of people carrying without permit actually goes up -- they are still carrying, just not bothering to try for a permit. As if they would have before...
You're saying that the number of guns available to criminals carrying guns will go up for the trivial reason that if everyone currently carrying a gun continues to regardless of more restrictive licensing, then more people will be classified as criminals. Well, OK, but I don't think that's what we're interested in: we're interested in the availability of guns to people who intend to commit violent crimes with them. You could argue that a further criminalization of gun ownership will cause some current peaceful gun owners to become violent criminals; perhaps, but I find it hard to believe that effect would be large.
Ignoring that effect and assuming the supply of violent gun-desiring criminals to be constant, the change in licensing is likely to restrict somewhat the availability of guns because, even if the criminals themselves don't care about the licensing, the people who they're stealing them from may, and hence the supply of stealable guns may be restricted.
Guns that are used consistently in crimes are generally stolen. The black market for guns would make them more valuable, but for people who are intent on carrying them, it would not be hard. You know supply & demand? As soon as guns are banned, demand skyrockets.
Woah, think about this carefully. You're arguing that a decrease in supply (banning guns), leads to an increase of prices, which you describe as an increase in "demand". This isn't the way the terms "supply" and "demand" are used in classical economics. The econ 101 argument would be that when supply decreases, people willing to pay more will do so and others will drop out of the market.
Depending on the inelasticity of the demand, it may be that a high proportion of the people buying guns may be willing to pay much higher prices (and/or to take much higher risks) in order to buy the same quantity of guns.
It's not obvious to me why that proprtion must be one hundred percent (equivalently, why the demand for guns must be completely inelastic).
Consider this. I work as a cashier, and luckily enough for me, I've never been the one with a gun pointed at me being told to hand over all the cash in the register.
But it's happened at my grocery store twice now.
Do you honestly think that guy with the gun pointed at that cashier's head had a permit for that firearm? Yeah, right.
The true meaning of the statement, "Make it a crime to carry a gun, and only criminals will carry guns" is not that people will become criminals by maintaining their guns. It means that the only people with guns will be those who intend to use them for criminal acts in the first place.
I'm not completely sure I understand what point you're trying to make, or how it relates to the above discussion of ccw revocations as an indicator of rates of violent crimes by ccw permit holders.
Are you arguing that whithholding of licenses will not reduce the availability of guns to criminals?
That's not clear to me. It may be true that people who hold up grocery stores don't usually have permits to carry guns, but that doesn't necessarily mean that withholding such permits would have no effect on the availability of such guns to such people. It seems likely, for example, that a significant number of such guns could be obtained from other criminals who make it their business to buy, sell, and/or steal guns that were originally obtained legitimately.
I haven't studied the problem at all, and can't quote any statistics; I'm just pointing out that if this is what you're trying to prove, then you need a more sophisticated argument than "guys who hold up grocery stores probably don't have ccw licenses."
*Any* restriction on the availability of firearms is likely to lead to *some* decrease in availability (equivalantly, increase in price) of guns to criminals. I imagine that the argument you'd want to make would be that this decrease in availability is outweighed by other advantages of more permissive licensing. It would require a more detailed study of the problem to determine how such an argument would work out.
Per the FBI, in 2002 there were 1,426,325 violent offenders in the United States. That was a rate of 494.6 per 100,000. Divide by 1,000, that's.4946/100, or about 0.5%. 2002 was, of course, near the bottom of a percipitous violent crime drop. In 2001, the number was 504.4/100,000 and 2002's number is a 12.9% decline from 1998.
That's the number of violent crimes reported to the FBI, which is likely more than the number of offenders, which is in turn more than the number of *convictions*, which is I believe what is relevant for revocations. The page you quote says the number of arrests in the same year was a little less than half the above number; presumably the number of convictions was less again. So the number of convictions of violent crime in that year was less than a quarter of a percent. Thus the fact that a given population has a rate of conviction of violent crimes of less than half a percent doesn't mark that population as unusual.
Since permits are revoked upon conviction for any violent crime, with or without a firearm, as well as many other reasons firearms crimes are merely a subset of that already small percentage. That percentage?.5%. One half of one percent.
.5% annually would be an extremely high percentage for violent crimes, so I don't think that number in itself is very useful. It'd be interesting to see more detailed statistics on those revocations.
In the US, how many police officers would have to die in a single incident to get that level of blanket news coverage? How many are shot and killed every week?
I don't know about the first question, but google will answer the second. E.g., from
http://www.cnn.com/US/9812/29/police.killed/ it looks like about 150 police die in the line of duty each year. From the same source, it looks like the causes are split in about half from traffic accidents and shootings, with other causes pretty negligible. So the answer to your question, it appears, is between 1 and 2 a week.
According to http://news.bbc.co.uk/1/hi/uk/2660885.stm, there were 14 police deaths in the UK in 2002, and this was considered high. I can't find any stastics on shooting deaths (though apparently it does happen: one example). The difference in population is a factor of about five, so the UK would seem to indeed have a comparatively lower rate.
Obviously, in a country of almost 300 million people, there's room for a lot of stupid and unfortunate things happen all the time. I don't know quite what to make of any of these statistics.
Sometimes I picture what it would be like if the current copyright laws were re-written so that ownership only existed for, oh, 15 years. Would a new set of industries pop up that release shows on various media formats?
For example, one company could be comitted to getting the content to you in the most inexpensive way possible. Another could be obsessed with video quality and extras (read: fanboys and their tv shows) and other such developments; they would charge a larger fee. Not to mention "fan sequals" and indy spinoffs.
Note that this already happens; just replace tv shows by, say, Shakespeare plays: go to any decent bookstore and look at their collection of Hamlet editions....
Default block anything below 1024 (in the appropriate direction, depending on the port), but let anyone explicitly request any given port to be opened, no questions asked. Quick signup on a web form, no long delay.
Apparently comcast zombies are one of the spammers' biggest tools, and remember that there are people who actually make their living sending spam. So they will have a *huge* incentive to work out ways around any blocking. This sounds like a pretty easy one to circumvent: the zombie software will just need a little extra smarts to talk to the firewall web form to open up port 25. Even if this requires a little monitoring to discover the user's comcast password, I bet this isn't very hard.
If they notice enough traffic to be of a concern (probably not only quantity, but it being sustained) they ring you and ask. IF you reply "Uhhh, what's SMTP?" they tell you you have a virus and send you to a page to get it diagnosed and fixed. If it's legit, they drop it.
I think part of the reason they haven't done this kind of manual intervention is that their business model depends on spending the absolute minimum of their employees' time on customers. People are expensive, more so people that understand what smtp is and can make the kind of judgement call you're asking them to, and $40--$50 a month doesn't buy very much of such a person's time.
This is more an acknowledgement that GNU/Linux is swimming in dangerous waters, and has enemies with money to burn.
On the other hand, the SCO case also showed that even a well-funded company given a year (so far) and great incentives to find intellectual property problems in the linux source has been unable to do so.
Not that it's bad for the linux developers to be careful and think ahead.
There are cryptographic methods available to the public at large (such as RC5 and PGP) that are proven to require more computing power than is theoretically possible in the universe.
No such proof exists. The best publicly known attacks may be computationally infeasible, and we may be given confidence based on our experience trying (and failing) to find more effective attacks. they are computationally infeasible to break. But noone has a proof of their strength, and it's always possible that dramatic advances in cryptanalysis may render an algorithm obsolete.
Okay... well, for the uninitiated, maybe I should explain that this is a religious issue of sorts among cycle freaks. The "Effective Cycling" crowd (led by John Forrester of MIT) are rather dogmatic in their insistence that bikes should simply behave like car traffic at all times. They're strongly biased against the idea of "bike lanes", because they might give the impression that bikes aren't *allowed* anywhere else.
Yes, it's only fair to say that there's a lot of disagreement on the subject of bike lanes (which I was trying to avoid by referring to "bike paths" that are separated from the street, which in general seem to have heavier evidence against them). On the other hand, I don't think it's fair to characterize Forrester's position as religious dogma, or to say that his only argument is one of appearances--they present more practical arguments which also need to be dealt with in any serious discussion of the issue.
Er, cyclists have a common-law right to use the road and always have. Motorists don't, they have to apply for a special permit called a license. And I don't know what the deal is in the US but in Europe it is illegal to ride a bike on the footpath.
In the US, the regulations regarding on bikes on sidewalks vary from locality to locality.
It is illegal and dangerous since pedestrians change their direction almost at random, an oncoming cyclist does not have time to avoid them.
Also, studies have found that sidewalk cyclists are more likely to have collisions with cars: the problem being that drivers know they have to scan to both sides for fast-moving traffic when crossing traffic lanes, but they don't do the same when they cross crosswalks.
Separated "bike paths" tend to have the same problems. The road really is where you want to be--and as someone who has shared the road with cars on daily commutes, I've found it works darned well in practice too. Just think of yourself as any other slow-moving vehicle (like a tractor (well, except when you have a sufficiently steep downhill...)) and it all works out.
Print out an alternate list of candidates, with your opponent swapped with an unlikely candidate. Stick it to the front of the voting machine. Anyone with 3 seconds unsupervised access to the machine can pull this off, and it may go unnoticed if it otherwise looks exactly like the original.
Interesting idea, but I think it would be hard to pull off (especially on a sufficiently large scale to have a reasonable chance of influencing an election) without detection. And it would be relatively easy to defeat if it proved necessary. (According to the article they limit votes to 5 a minute, which would leave plenty of time for poll workers to check the machines periodically.)
Oh, and one more recommendation: make sure that you take at least one real theoretical math class; you'll know you're in the right class if 90% of the homework problems are of the form "prove the following statement". The subject matter doesn't even matter that much; abstract algebra or real analysis are often the first courses where math departments start expecting students to understand proofs, but ask a professor in your local math department and they'll be able to recommend the right course.
A good program is *exactly* like a good proof: clear, concise, and convincing.
Perhaps the most related, useful, but most underappreciated related discipline is statistics. Of all the non-CS classes I took, stats is the most relevant to my day-to-day life. For example, doing analysis of performance and tuning software system, I often see people use bogus statistical analysis, and making mistakes based on those results. Even if your curriculum doesn't require it, I would highly recommend taking a stats class or two.
Amen. Mod parent up some more. A little statistics and experimental design will go a long way.
My list of math courses would be:
A couple semester's worth of applied statistics.
A couple semester's worth of combinatorics/discrete mathematics. Graham, Knuth, and Patashnik's "Concrete Mathematics" is my personal favorite for this.
The usual year or two of calculus, including a bit of multivariable calculus and at least a semester of linear algebra.
And if you're really into it, throw in some algebra and number theory, just because it's fun, and has some amusing computer science applications. Oh, and some differential equations never hurt anyone. Just get your BA in math and then MS in computer science....
You must rent a cardboard box and eat scraps from a garbage bin and steal cable and have a walkie talkie for a phone and drive a scooter to have a sizeable chunk left from that pay.
Scooter: huh? Walk or ride a bike; both are free and better for you. Rent, so you're not stuck someplace out of range if your job moves.
eat scraps from a garbage bin: You can have excellent food for under $200 a month easily if you're willing to cook instead of eating out all the time. It's better for you, too.
walkie talkie for a phone: basic phone service where I live, with taxes and all, is under $25 a month. Even better, split it with someone. You don't need the stupid cell phone. Ditch the cable (and the TV) while you're at it and learn to like the local library.
rent a cardboard box: At $2600 a year (assume 40 hours a week), you're not paying much in taxes; the above expenses plus utilities and random stuff needn't be more than $3000 or so. That leaves over $1500 you could pay in rent if you had to. That should get you something within walking distance of work even in most cities, better if you can split the place with someone. Stick the extra in an IRA.
What the hell was wrong with him using his affiliate link?
When I see such a link I wonder whether the person posting it is making an honest recommendation, or whether they just post links to books on amazon everywhere all the time just to make a few bucks.
"will always be dominated by a couple of very centrist candidates"
This isn't really true in the U.S. lately.....There really is no liberal political party in the U.S. any more. The right has, over the last 25 years, successfully tarred the media as liberal and all liberals as dangerous.
By "centrist" I meant "center of the spectrum of electable candidates", not "center of the spectrum of opinions that might be held by sane, informed people". I agree that the former is well to the right of the latter these days.
That's why I encourage people to vote for minor parties - if nobody does because "they'll never win" then we will always be stuck with the two party system.
And if one of the really does win big, then you'll just end up with another two-party system, with different parties. This has happened before. It's just the way the US government (like other governments dominated by two parties) is structured--there's no advantage to coming in second in an election (unlike in a system with at-large candidates that can form coalitions), so there's no benefit to voting for someone that doesn't have a reasonable chance of winning, so noone sane is going to throw a lot of resources behind a clear third-place candidate. For this reason presidential elections in particular will always be dominated by a couple of very centrist candidates.
So if you really want to see viable third parties then you need a constitutional amendment, or two or three.
Personally I'd rather expend my political energy elsewhere....
Slashdot Mirror
And it should be pointed out that even those of us whose preferred mode of configuration is something like that still depend on certain conventions, which can do a great deal to make the software more or less usable.
E.g., the configuration files should be located at a path in etc with "smb" or "samba" somewhere in the name, documentation should be deposited in the standard places, the configuration file should use a format that's also used by other programs instead of one you just made up, etc., etc.
Similarly gui configurators should be findable in some standard location, should use widgets users have seen before, should label options clearly, etc.
People focus so much on the newbie/sysadmin and gui/commandline dichotomy's that they forget that the problems faced by both are very similar.
--Bruce Fields
And how many people died in this incident? And how many people pass through downtown Portland in the average day? If you were really worried about your chances of being killed while walking through downtown Portland, surely you'd be better off, say, working on your street-crossing technique.
Linux? I thought it was the city.... Portland rocks. I'd be back there in a minute if I had a good excuse.
--Bruce Fields
Try a half-million. Wikepedia was the first reference I found for this on google; perhaps someone else can find something more authoritative.
--Bruce Fields
Huh? I spent lots of time wandering around downtown Portland at all hours of the night and day (and lived at 16th and Yamhill for a year) in the early ninetees, and can't think of anything that would prompt this comment. It didn't look much different when I visted there last summer.
If you were going to feel weird about any area, OK, maybe just north of Burnside is a little seedy, but I never had any actual trouble there either.
--Bruce Fields
You're saying that the number of guns available to criminals carrying guns will go up for the trivial reason that if everyone currently carrying a gun continues to regardless of more restrictive licensing, then more people will be classified as criminals. Well, OK, but I don't think that's what we're interested in: we're interested in the availability of guns to people who intend to commit violent crimes with them. You could argue that a further criminalization of gun ownership will cause some current peaceful gun owners to become violent criminals; perhaps, but I find it hard to believe that effect would be large.
Ignoring that effect and assuming the supply of violent gun-desiring criminals to be constant, the change in licensing is likely to restrict somewhat the availability of guns because, even if the criminals themselves don't care about the licensing, the people who they're stealing them from may, and hence the supply of stealable guns may be restricted.
Woah, think about this carefully. You're arguing that a decrease in supply (banning guns), leads to an increase of prices, which you describe as an increase in "demand". This isn't the way the terms "supply" and "demand" are used in classical economics. The econ 101 argument would be that when supply decreases, people willing to pay more will do so and others will drop out of the market.Depending on the inelasticity of the demand, it may be that a high proportion of the people buying guns may be willing to pay much higher prices (and/or to take much higher risks) in order to buy the same quantity of guns.
It's not obvious to me why that proprtion must be one hundred percent (equivalently, why the demand for guns must be completely inelastic).
I'm not completely sure I understand what point you're trying to make, or how it relates to the above discussion of ccw revocations as an indicator of rates of violent crimes by ccw permit holders.
Are you arguing that whithholding of licenses will not reduce the availability of guns to criminals?
That's not clear to me. It may be true that people who hold up grocery stores don't usually have permits to carry guns, but that doesn't necessarily mean that withholding such permits would have no effect on the availability of such guns to such people. It seems likely, for example, that a significant number of such guns could be obtained from other criminals who make it their business to buy, sell, and/or steal guns that were originally obtained legitimately.
I haven't studied the problem at all, and can't quote any statistics; I'm just pointing out that if this is what you're trying to prove, then you need a more sophisticated argument than "guys who hold up grocery stores probably don't have ccw licenses."
*Any* restriction on the availability of firearms is likely to lead to *some* decrease in availability (equivalantly, increase in price) of guns to criminals. I imagine that the argument you'd want to make would be that this decrease in availability is outweighed by other advantages of more permissive licensing. It would require a more detailed study of the problem to determine how such an argument would work out.
--Bruce Fields
That's the number of violent crimes reported to the FBI, which is likely more than the number of offenders, which is in turn more than the number of *convictions*, which is I believe what is relevant for revocations. The page you quote says the number of arrests in the same year was a little less than half the above number; presumably the number of convictions was less again. So the number of convictions of violent crime in that year was less than a quarter of a percent. Thus the fact that a given population has a rate of conviction of violent crimes of less than half a percent doesn't mark that population as unusual.
.5% annually would be an extremely high percentage for violent crimes, so I don't think that number in itself is very useful. It'd be interesting to see more detailed statistics on those revocations.
I don't know about the first question, but google will answer the second. E.g., from http://www.cnn.com/US/9812/29/police.killed/ it looks like about 150 police die in the line of duty each year. From the same source, it looks like the causes are split in about half from traffic accidents and shootings, with other causes pretty negligible. So the answer to your question, it appears, is between 1 and 2 a week.
According to http://news.bbc.co.uk/1/hi/uk/2660885.stm, there were 14 police deaths in the UK in 2002, and this was considered high. I can't find any stastics on shooting deaths (though apparently it does happen: one example). The difference in population is a factor of about five, so the UK would seem to indeed have a comparatively lower rate.
Obviously, in a country of almost 300 million people, there's room for a lot of stupid and unfortunate things happen all the time. I don't know quite what to make of any of these statistics.
--Bruce Fields
Note that this already happens; just replace tv shows by, say, Shakespeare plays: go to any decent bookstore and look at their collection of Hamlet editions....
--Bruce Fields
Apparently comcast zombies are one of the spammers' biggest tools, and remember that there are people who actually make their living sending spam. So they will have a *huge* incentive to work out ways around any blocking. This sounds like a pretty easy one to circumvent: the zombie software will just need a little extra smarts to talk to the firewall web form to open up port 25. Even if this requires a little monitoring to discover the user's comcast password, I bet this isn't very hard.
--Bruce Fields
I think part of the reason they haven't done this kind of manual intervention is that their business model depends on spending the absolute minimum of their employees' time on customers. People are expensive, more so people that understand what smtp is and can make the kind of judgement call you're asking them to, and $40--$50 a month doesn't buy very much of such a person's time.
--Bruce Fields
On the other hand, the SCO case also showed that even a well-funded company given a year (so far) and great incentives to find intellectual property problems in the linux source has been unable to do so.
Not that it's bad for the linux developers to be careful and think ahead.
--Bruce Fields
No such proof exists. The best publicly known attacks may be computationally infeasible, and we may be given confidence based on our experience trying (and failing) to find more effective attacks. they are computationally infeasible to break. But noone has a proof of their strength, and it's always possible that dramatic advances in cryptanalysis may render an algorithm obsolete.
--Bruce Fields
Yes, it's only fair to say that there's a lot of disagreement on the subject of bike lanes (which I was trying to avoid by referring to "bike paths" that are separated from the street, which in general seem to have heavier evidence against them). On the other hand, I don't think it's fair to characterize Forrester's position as religious dogma, or to say that his only argument is one of appearances--they present more practical arguments which also need to be dealt with in any serious discussion of the issue.
--Bruce Fields
In the US, the regulations regarding on bikes on sidewalks vary from locality to locality.
Also, studies have found that sidewalk cyclists are more likely to have collisions with cars: the problem being that drivers know they have to scan to both sides for fast-moving traffic when crossing traffic lanes, but they don't do the same when they cross crosswalks.
Separated "bike paths" tend to have the same problems. The road really is where you want to be--and as someone who has shared the road with cars on daily commutes, I've found it works darned well in practice too. Just think of yourself as any other slow-moving vehicle (like a tractor (well, except when you have a sufficiently steep downhill...)) and it all works out.
--Bruce Fields
Interesting idea, but I think it would be hard to pull off (especially on a sufficiently large scale to have a reasonable chance of influencing an election) without detection. And it would be relatively easy to defeat if it proved necessary. (According to the article they limit votes to 5 a minute, which would leave plenty of time for poll workers to check the machines periodically.)
--Bruce Fields
Oh, and one more recommendation: make sure that you take at least one real theoretical math class; you'll know you're in the right class if 90% of the homework problems are of the form "prove the following statement". The subject matter doesn't even matter that much; abstract algebra or real analysis are often the first courses where math departments start expecting students to understand proofs, but ask a professor in your local math department and they'll be able to recommend the right course.
A good program is *exactly* like a good proof: clear, concise, and convincing.
--Bruce Fields
Amen. Mod parent up some more. A little statistics and experimental design will go a long way.
My list of math courses would be:
And if you're really into it, throw in some algebra and number theory, just because it's fun, and has some amusing computer science applications. Oh, and some differential equations never hurt anyone. Just get your BA in math and then MS in computer science....
--Bruce Fields
Exactly how do you think turning on the "tainted" flag for binary drivers constitutes "fanaticism"?
How am I a "fanatic" if I don't want to help debug a kernel that includes code that the author won't share with me?
--Bruce Fields
When I see such a link I wonder whether the person posting it is making an honest recommendation, or whether they just post links to books on amazon everywhere all the time just to make a few bucks.
--b.
By "centrist" I meant "center of the spectrum of electable candidates", not "center of the spectrum of opinions that might be held by sane, informed people". I agree that the former is well to the right of the latter these days.
--Bruce Fields
And if one of the really does win big, then you'll just end up with another two-party system, with different parties. This has happened before. It's just the way the US government (like other governments dominated by two parties) is structured--there's no advantage to coming in second in an election (unlike in a system with at-large candidates that can form coalitions), so there's no benefit to voting for someone that doesn't have a reasonable chance of winning, so noone sane is going to throw a lot of resources behind a clear third-place candidate. For this reason presidential elections in particular will always be dominated by a couple of very centrist candidates.
So if you really want to see viable third parties then you need a constitutional amendment, or two or three.
Personally I'd rather expend my political energy elsewhere....
--Bruce Fields