No, there isn't. In common parlance, TCP/IP means all kinds of network protocols running on top of IP, such as TCP, UDP, ICMP, and others. If you specifically mean only TCP, then you just say TCP, not TCP/IP.
If you can't find a coherent argument for why it's in your companys interest to release the source code under the GPL, then there is probably little reason to do it...
Then again, unless your company is in the business of selling "intranet/extranet portal framework"s then it shouldn't hurt much.
Apart from GNU ideology, the decision boils down to:
Will it create good relations to someone the company cares about?
Is the source code embarassing to show to others?
Will the company actually benefit from other users sending patches? (and no, this isn't as clearcut as open source ideologes claim, see the question below...)
Will this take significant amounts of your time that would be better spent doing other things that benefit the company more?
If you can find reasonable answers to those questions, a reasonable boss will make a reasonable decision:-)
The "better" spam filters described by Graham are already getting pretty common in decent mail user agents. And yes, bayesian filtering works well.
However, it will not make spam unprofitable. To make it unprofitable, the costs of sending spam must be higher than the money you get from it. So in some way, we need to increase the costs of sending spam, or reduce profits.
The cost of sending spam is essentially zero. Sure, you may have to switch ISP once in a while, register some new domains, invest in some CDs with email-addresses, buy some software or consultants to infect machines, etc... But it really doesn't matter. Even with todays hostility towards spammers, the cost is still essentially zero.
where really_low_fraction is the number of idiots who fall for your scam.
Bayesian filtering doesn't address either costs or profits. It does not make sending spam more expensive, and it does not change the some_really_low_fraction, because the idiots who respond to spam wouldn't be using bayesian filtering anyway.
So Bayesian filtering is nice for the end-users who just want to get through their mail, but it doesn't really help solve the problem of making spam unprofitable.
Didn't Torvalds once say something along the line that 'perfect is the enemy of good' when criticizing BSD? Is he moving away from 'good-enough' with lots of features constantly coming out, towards a more BSD-esque, move along slowly with stable-code philosophy?
Linus says a lot of things. It seems to me that he is just using the scientific approach, and trying new ways of doing stuff to see what works best. Some ideas are good, others are bad. But if you never change your process, you'll never find out.
These changes in the process make a lot of people scream whenever they happen. That's because people doesn't like change. Even now, people are screaming about breaking the odd/even process (which didn't work too well), even though the 2.6 process has worked much better. If the 2.6.13 process isn't even better, Linus will scrap it and try something else (such as going back to the old 2.6 process, or the 2.6.x.y process, or something else new, or whatever).
Stay calm! The world isn't going to end! All these changes mostly affects kernel developers, and even then, mostly those in the "inner circle". Your redhat/ubuntu/suse/whatever will still work just fine.
Not to be a troll or anything, but why is everyone freaking out so much over google maps?
Probably because they are americans. This is a map showing Oslo (the capital of Norway), and surrounding areas. Note the complete lack of any roads, names, or any useful information at all. If you zoom out, you will find out that google knows that Norway is a country, but it doesn't even know the name of our capital.
I don't think you'll find many open source testing tools. Usually open source software is tested by its users. While that is hopelessly inadequate if you have 5 users, and intend to charge them big bucks, it tends to work well in the open source world.
That's entirely possible, but its chemical properties will remove it long before it reaches that height. You see, our atmosphere is a bit atypical in that it contains oxygene, which pretty much reacts with anything, and certainly with hydrogen.
To run something in Linux that can potentially damage the system you need to log in as root. To run a virus you need to submit root password which is pretty different from what happens in Windows (by the way can you run Windows as restricted user? Many programs just refuse to work, I think that restricted user account is useless, most of the people I know run Windows as Administrator, only that and makes a big difference.)
Well, but then again, this isn't about desktops. The article was about somebody managing a shitload of windows and linux servers. And I assume that meant that they (a) knew what they were doing (b) had a reasonable security infrastructure (c) didn't need to worry about desktops.
Given the premises, I actually believe them.
Actually, I would believe them, if they talked about desktops too. A reasonable corporate environment with windows desktops will be locked down to the extreme. A reasonable corporate environment with linux desktops will more likely run some standard distribution, and be fairly lax in security. They should score similarly.
As for home users who don't have experts to secure their computers, there is still another story.
But A: it could have wireless, and B: even non idiots will possibly connect it just for a little while - I mean you have this nice new computer just waiting to be used, let's browse just a little. And only then wipe it clean.
Because it's based on pretty sound knowledge-base-engineering. That doesn't necessarily mean it's good, or even useful, or that first-order logic is a reasonable medium to teach computers common sense. All it means is that it's among the best we have for this kind of stuff.
If you need a common-sense reasoning engine, combined with common-sense facts, mindpixels stuff won't help you at all. It's just a stupid program any child could write, and a bunch of useless sentences with a probability value. On the other hand; CYC might help. The obvious alternative would be to build one yourself. But even then, you would be a fool to overlook parts of CYC that would be useful for you.
While you are perfectly justified in doubting whether CYC is useful for anything at all, creating something significantly better, would take both a lot of time, and a spark of genius.
I know one guy who worked on the project who thought the whole thing was a scam. I'm yet to hear a report of Cyc actually doing anything interesting or useful.
Given that they still can't put a reasonable "success story" on the frontpage for the commercial version, I tend to agree that they should be pretty careful in what they claim. But so they are (apart from claiming to be far more complete than they actually are). In comparison, mindpixel spams every newsgroup and discussion board he sees, trying to be the messias of the AI revolution.
And Lenat's previous projects (eg. AM, Eurisko) seem very suspect to me. If they really could do what they did at the time (late 70s, early 80s) we should see some amazing stuff today along those lines. We don't.
We now can have a 5-year anniversary of bashing a completely useless project brought to you by some internet cook who thinks he has "solved" AI by writing a program that even a 5-year old would understand is useless.
If you want a real database of "common-sense" knowledge, you should check out CYC instead. It might be harder to do it that way, but it sure pays off if you actually want to use it for something beyond spamming usenet groups and slashdot.
2. If your laptop is stolen, by the time it manages to report it to the police, it will be too late.
Too late for what? For recovery? No. For prosecution of the thief? Probably. For prosecution of the moron who bought it and knew it was stolen? No.
It's unlikely anyone but the last buyer will even attempt to connect it to the Internet. So whether the police uses 12 weeks or 4 months to get to him doesn't matter much, they will still find the laptop, and someone to put in jail.
On the other hand, the mechanism only works on idiots. If I were to buy a stolen laptop (not that I'm into that kind of thing anyway), I would of course wipe it clean, just as I do with any other new or used computer that gets into my hands...
I think you forgot the part that said win32 firewalls. These firewalls block programs from sending outgoing requests. I.e. internet explorer is allowed to connect, but not phonehome.exe. Ports doesn't matter much.
There are tricks to work around this, using dll's or explorer plugins, but some win32 firewalls can deal with that with some sofistication too (although it's more painful for the user to configure it right).
Then again, I bet people who use stolen laptops tend to wipe the disk clean, unless they're into industrial espionage, in which case they would probably take a lot of precautions anyway (and certainly not connect it to the Internet). A better solution would be hardware-based.
qsort() is a very well-understood algorithm that has been highly optimized.
Bzzt, wrong. qsort() is NOT, I repeat, NOT, the same as quicksort.
According to unix manpages: "The qsort subroutine sorts a table of data in place. It uses the quicker-sort algorithm.
Note that "quicker-sort" is not the same as "quicksort". In fact, I doubt anyone really knows what the hell "quicker sort" is, anyway. My guess is that it might be a name just invented to explain away the q, once some standardization committee decided to not have implementation details in the spec (assuming qsort() really meant quicksort some time in the past).
So, qsort() is not a very well understood algorithm. And I seriously doubt it has been highly optimized, as there is no point in it.
You see, for every comparison qsort() does, it has to call a function through a function pointer. That is pretty brain-damaged for the typical case, where the comparison is typically done in just one or two CPU opcodes. Spending lots of effort optimizing something that is a lost race anyway, is not a wise use of time for compiler vendors. For this reason, qsort() is more likely to be a slightly naïve implementation of quicksort.
Not including it in the benchmarks would have been a sign that some smoke and mirrors were involved.
Actually, I would say the opposite. Anyone caring even a little bit about performance would run away screaming if someone suggested they'd use qsort(). Benchmarking against something as sluggish as that IS a sign that some smoke and mirrors were involved.
If the MSVC++ library had anything faster, people would be using it.
The obvious answer is to make a qsort kernel module and get it out of user space!
Yeah, that would be really neat. Not only do you still have to call a function for each comparison. But *each* and *every one* of those function calls will need to go across the kernel-userspace boundary. My oh my, you are clever (not!)
The perfect OS for blind people is any operating system using the command line, such as the many unix-variants, DOS, VMS, or anything else except windows or mac. And the perfect text-editor is line-oriented; e.g. ed) (or DOS's edlin:-)
The problem isn't that such OSes or application software don't exist. The problem is that most tasks people do with PCs, is not something that is suitable for blind people.
Case in point: most of todays office or "productivity" software is WYSIWYG. That is pretty silly for blind people. Having the perfect web-browser for blind people (lynx, links, or perhaps even better: w3m inside emacspeak) doesn't help much when web-sites use image-maps or javascript-menus for navigation, or drown the user in meaningless hyperlinks, taking hours to tab through.
So how do you design the perfect OS for blind people? You don't. It's already here, and it's called linux (or ***bsd). And if blind people still want to use PCs for the same things as most of the seeing people use them for, they will have to live with some compromises. And maybe even use windows, because that's where you will find the best commercial accessability software...
Sorry...
(But then again, you shouldn't ask slashdot how to design an OS for blind people, you should ask blind people...)
In fact, how much more likely is it that a flash-of-insight invention will be forgotten, or merely jotted down, because of the costs involved in getting it to market?
I have no idea. But if someone would sell more paper-driers, or could charge more for the ones they already sold, simply by using gravity better, then I believe our (capitalistic) system tends to work as intended. Either that, or people would "mod" their paper-driers, and manufacturers would eventually wisen up (or maybe a whole separate paper-drier-mod industry would start to flourish...)
The idea that something should be "forgotten" simply because it was not patented, is ridiculous. Tell me, when was the last time you curiously browsed patent submissions, looking for good ideas to use? My guess is never. Instead, like everybody else, you get your ideas from clever people, books, conferences, periodicals, internet, industrial espionage, etc..., and certainly not not from patent applications.
If an idea gets forgotten, it's usually because it wasn't that good anyway. Giving it an official patent number won't make it any better.
If the idea is good, and as you say, is forgotten because it wasn't practical/economical at the time it was discovered, it can either be written down elsewhere, or (more likely) rediscovered later by some other person. (It was after all a "flash-of-insight" idea we spoke about, which will no doubt be discovered again and again untill it gets practical)
But what if all your competitors have the same bug, and you have somehow managed to avoid it? I would still consider that a feature. And having a better design that avoids bugs or misfeatures, is definitely a feature.
But no one really thought of the problem in those terms (or just wasn't thinking of the problem at all), and what ended up a pretty obvious solution, completely changed how quickly people could get their paper making done.
Yes, that's true. But why do we have patents? Is it to reward inventors? Or is it to encourage inventors (or others) to invent more?
The idea of using gravity, might have needed a spark of creativity in someones head, at one time. And it is a good idea.
But it's not the kind of idea that costs thousands of dollars to develop, that others can copy for nothing. It is the kind of idea, that someone would think of sooner or later, and as soon as someone thinks of it, everyone benefits. That is, unless it got patented.
As with most software patents, patenting this kind of stuff, does not benefit society at large. Just making that random person richer, will not encourage anyone to invent more.
Yes, if everything is running in true 64bit mode and is using the optimizations, it will run faster. However, many optimizations must be implemented at a very low level and many programs are not utilizing the full potential.
Huh? Taking care of "low-level" stuiff, is is exactly what a compiler is doing. If you recompile for amd64, it will usually get faster (unless it gets slower due to cache issues).
One such program that is, is MPlayer, a good portion of the code is implemented in Assembly specifically taking advantage of the 64bit architecture. But, if you install a x86 version of Linux or Windows on this hardware, you won't actually see an increase in speed.
Huh? You don't think mplayer is at least as good optimized on x86 as amd64? That would be pretty strange, as x86 has been it's main target all along. Besides, unless you're running on x86, you can't use the native windows codecs, making mplayer somewhat less useful.
Not having something is not a feature, no matter how desperately you'd like it to be.
At the moment, I'm not having diarrhea. I consider that a feature, and not a bug. (STREAMS under un*x-like operating systems is comparable to diarrhea in this respect...)
Slashdot Mirror
No, jonadab was correct. foldoc link
No, there isn't. In common parlance, TCP/IP means all kinds of network protocols running on top of IP, such as TCP, UDP, ICMP, and others. If you specifically mean only TCP, then you just say TCP, not TCP/IP.
If you can't find a coherent argument for why it's in your companys interest to release the source code under the GPL, then there is probably little reason to do it...
Then again, unless your company is in the business of selling "intranet/extranet portal framework"s then it shouldn't hurt much.
Apart from GNU ideology, the decision boils down to:
If you can find reasonable answers to those questions, a reasonable boss will make a reasonable decision :-)
However, it will not make spam unprofitable. To make it unprofitable, the costs of sending spam must be higher than the money you get from it. So in some way, we need to increase the costs of sending spam, or reduce profits.
The cost of sending spam is essentially zero. Sure, you may have to switch ISP once in a while, register some new domains, invest in some CDs with email-addresses, buy some software or consultants to infect machines, etc... But it really doesn't matter. Even with todays hostility towards spammers, the cost is still essentially zero.
The profits of spam is:
- price_of_whatever_you_sell * number_of_email_addresses * some_really_low_fraction
where really_low_fraction is the number of idiots who fall for your scam.Bayesian filtering doesn't address either costs or profits. It does not make sending spam more expensive, and it does not change the some_really_low_fraction, because the idiots who respond to spam wouldn't be using bayesian filtering anyway.
So Bayesian filtering is nice for the end-users who just want to get through their mail, but it doesn't really help solve the problem of making spam unprofitable.
Linus says a lot of things. It seems to me that he is just using the scientific approach, and trying new ways of doing stuff to see what works best. Some ideas are good, others are bad. But if you never change your process, you'll never find out.
These changes in the process make a lot of people scream whenever they happen. That's because people doesn't like change. Even now, people are screaming about breaking the odd/even process (which didn't work too well), even though the 2.6 process has worked much better. If the 2.6.13 process isn't even better, Linus will scrap it and try something else (such as going back to the old 2.6 process, or the 2.6.x.y process, or something else new, or whatever).
Stay calm! The world isn't going to end! All these changes mostly affects kernel developers, and even then, mostly those in the "inner circle". Your redhat/ubuntu/suse/whatever will still work just fine.
Probably because they are americans. This is a map showing Oslo (the capital of Norway), and surrounding areas. Note the complete lack of any roads, names, or any useful information at all. If you zoom out, you will find out that google knows that Norway is a country, but it doesn't even know the name of our capital.
Well, back to http://finn.no/kart/ for me...
Yeah, but we already have plenty of JVMs to choose from. The problem is the rest of the JDK.
Does a phone with a backup battery count?
I don't think you'll find many open source testing tools. Usually open source software is tested by its users. While that is hopelessly inadequate if you have 5 users, and intend to charge them big bucks, it tends to work well in the open source world.
That's entirely possible, but its chemical properties will remove it long before it reaches that height. You see, our atmosphere is a bit atypical in that it contains oxygene, which pretty much reacts with anything, and certainly with hydrogen.
You just did :-)
Well, but then again, this isn't about desktops. The article was about somebody managing a shitload of windows and linux servers. And I assume that meant that they (a) knew what they were doing (b) had a reasonable security infrastructure (c) didn't need to worry about desktops.
Given the premises, I actually believe them.
Actually, I would believe them, if they talked about desktops too. A reasonable corporate environment with windows desktops will be locked down to the extreme. A reasonable corporate environment with linux desktops will more likely run some standard distribution, and be fairly lax in security. They should score similarly.
As for home users who don't have experts to secure their computers, there is still another story.
Good point.
Because it's based on pretty sound knowledge-base-engineering. That doesn't necessarily mean it's good, or even useful, or that first-order logic is a reasonable medium to teach computers common sense. All it means is that it's among the best we have for this kind of stuff.
If you need a common-sense reasoning engine, combined with common-sense facts, mindpixels stuff won't help you at all. It's just a stupid program any child could write, and a bunch of useless sentences with a probability value. On the other hand; CYC might help. The obvious alternative would be to build one yourself. But even then, you would be a fool to overlook parts of CYC that would be useful for you.
While you are perfectly justified in doubting whether CYC is useful for anything at all, creating something significantly better, would take both a lot of time, and a spark of genius.
I know one guy who worked on the project who thought the whole thing was a scam. I'm yet to hear a report of Cyc actually doing anything interesting or useful.
Given that they still can't put a reasonable "success story" on the frontpage for the commercial version, I tend to agree that they should be pretty careful in what they claim. But so they are (apart from claiming to be far more complete than they actually are). In comparison, mindpixel spams every newsgroup and discussion board he sees, trying to be the messias of the AI revolution.
And Lenat's previous projects (eg. AM, Eurisko) seem very suspect to me. If they really could do what they did at the time (late 70s, early 80s) we should see some amazing stuff today along those lines. We don't.
Yes, AI is still far from "solved".
If you want a real database of "common-sense" knowledge, you should check out CYC instead. It might be harder to do it that way, but it sure pays off if you actually want to use it for something beyond spamming usenet groups and slashdot.
Too late for what? For recovery? No. For prosecution of the thief? Probably. For prosecution of the moron who bought it and knew it was stolen? No.
It's unlikely anyone but the last buyer will even attempt to connect it to the Internet. So whether the police uses 12 weeks or 4 months to get to him doesn't matter much, they will still find the laptop, and someone to put in jail.
On the other hand, the mechanism only works on idiots. If I were to buy a stolen laptop (not that I'm into that kind of thing anyway), I would of course wipe it clean, just as I do with any other new or used computer that gets into my hands...
There are tricks to work around this, using dll's or explorer plugins, but some win32 firewalls can deal with that with some sofistication too (although it's more painful for the user to configure it right).
Then again, I bet people who use stolen laptops tend to wipe the disk clean, unless they're into industrial espionage, in which case they would probably take a lot of precautions anyway (and certainly not connect it to the Internet). A better solution would be hardware-based.
Bzzt, wrong. qsort() is NOT, I repeat, NOT, the same as quicksort.
According to unix manpages: "The qsort subroutine sorts a table of data in place. It uses the quicker-sort algorithm.
Note that "quicker-sort" is not the same as "quicksort". In fact, I doubt anyone really knows what the hell "quicker sort" is, anyway. My guess is that it might be a name just invented to explain away the q, once some standardization committee decided to not have implementation details in the spec (assuming qsort() really meant quicksort some time in the past).
So, qsort() is not a very well understood algorithm. And I seriously doubt it has been highly optimized, as there is no point in it.
You see, for every comparison qsort() does, it has to call a function through a function pointer. That is pretty brain-damaged for the typical case, where the comparison is typically done in just one or two CPU opcodes. Spending lots of effort optimizing something that is a lost race anyway, is not a wise use of time for compiler vendors. For this reason, qsort() is more likely to be a slightly naïve implementation of quicksort.
Not including it in the benchmarks would have been a sign that some smoke and mirrors were involved.
Actually, I would say the opposite. Anyone caring even a little bit about performance would run away screaming if someone suggested they'd use qsort(). Benchmarking against something as sluggish as that IS a sign that some smoke and mirrors were involved.
If the MSVC++ library had anything faster, people would be using it.
Doesn't MSVC++ have std::sort() ?
Yeah, that would be really neat. Not only do you still have to call a function for each comparison. But *each* and *every one* of those function calls will need to go across the kernel-userspace boundary. My oh my, you are clever (not!)
The problem isn't that such OSes or application software don't exist. The problem is that most tasks people do with PCs, is not something that is suitable for blind people.
Case in point: most of todays office or "productivity" software is WYSIWYG. That is pretty silly for blind people. Having the perfect web-browser for blind people (lynx, links, or perhaps even better: w3m inside emacspeak) doesn't help much when web-sites use image-maps or javascript-menus for navigation, or drown the user in meaningless hyperlinks, taking hours to tab through.
So how do you design the perfect OS for blind people? You don't. It's already here, and it's called linux (or ***bsd). And if blind people still want to use PCs for the same things as most of the seeing people use them for, they will have to live with some compromises. And maybe even use windows, because that's where you will find the best commercial accessability software...
Sorry...
(But then again, you shouldn't ask slashdot how to design an OS for blind people, you should ask blind people...)
I have no idea. But if someone would sell more paper-driers, or could charge more for the ones they already sold, simply by using gravity better, then I believe our (capitalistic) system tends to work as intended. Either that, or people would "mod" their paper-driers, and manufacturers would eventually wisen up (or maybe a whole separate paper-drier-mod industry would start to flourish...)
The idea that something should be "forgotten" simply because it was not patented, is ridiculous. Tell me, when was the last time you curiously browsed patent submissions, looking for good ideas to use? My guess is never. Instead, like everybody else, you get your ideas from clever people, books, conferences, periodicals, internet, industrial espionage, etc..., and certainly not not from patent applications.
If an idea gets forgotten, it's usually because it wasn't that good anyway. Giving it an official patent number won't make it any better.
If the idea is good, and as you say, is forgotten because it wasn't practical/economical at the time it was discovered, it can either be written down elsewhere, or (more likely) rediscovered later by some other person. (It was after all a "flash-of-insight" idea we spoke about, which will no doubt be discovered again and again untill it gets practical)
But what if all your competitors have the same bug, and you have somehow managed to avoid it? I would still consider that a feature. And having a better design that avoids bugs or misfeatures, is definitely a feature.
Yes, that's true. But why do we have patents? Is it to reward inventors? Or is it to encourage inventors (or others) to invent more?
The idea of using gravity, might have needed a spark of creativity in someones head, at one time. And it is a good idea. But it's not the kind of idea that costs thousands of dollars to develop, that others can copy for nothing. It is the kind of idea, that someone would think of sooner or later, and as soon as someone thinks of it, everyone benefits. That is, unless it got patented.
As with most software patents, patenting this kind of stuff, does not benefit society at large. Just making that random person richer, will not encourage anyone to invent more.
Huh? Taking care of "low-level" stuiff, is is exactly what a compiler is doing. If you recompile for amd64, it will usually get faster (unless it gets slower due to cache issues).
One such program that is, is MPlayer, a good portion of the code is implemented in Assembly specifically taking advantage of the 64bit architecture. But, if you install a x86 version of Linux or Windows on this hardware, you won't actually see an increase in speed.
Huh? You don't think mplayer is at least as good optimized on x86 as amd64? That would be pretty strange, as x86 has been it's main target all along. Besides, unless you're running on x86, you can't use the native windows codecs, making mplayer somewhat less useful.
At the moment, I'm not having diarrhea. I consider that a feature, and not a bug. (STREAMS under un*x-like operating systems is comparable to diarrhea in this respect...)