He has his reasons, and extremely often he is right. (I think he's wrong about that proposal about the next gen email method tho, not that it matters).
Yeah he is arrogant and often obnoxious but I actually regard this as a feature in this case.
Why? He must have pissed off thousands of people by now (many of them fairly smart too). The way these people can get back at him where it hurts him is to find a security flaw in his program and rub his nose in it.
The fact that I haven't seen any qmail and djbdns exploits (ala remote root or remote user ) despite all that is gives me more confidence in the security of his software.:)
Hmm, some RSI-like symptoms are actually due to nerve problems. Maybe in this case the nerves are weakened and so are more sensitive to just a bit of squishing.
Did you actually test MAPS/ORBS to see how accurate they are? Is your mailserver configured to do further processing on mails from IPs listed on MAPS/ORBS?
At home I regularly get 100-200+ spams a day. Spambayes screens most of them with minimal false positives (there were a few false positives during the initial training period, zero since then - I'm still testing it ). Even my personal naive keyword filtering screened out plenty of spam with very minimal false positives.
Disclaimer: My company resells Puremessage - it can be configured to take into consideration a listing on DNS blacklists. Taking into consideration means adding spamminess points, but still going through with the rest of the spamminess and hamminess tests.
Personally the MAPS/ORBS stuff is pretty shoddy. In the old days they even let you download lists of open relays. Now that must have been pretty useful for spammers and some other folk... To me that indicated to me how the MAPS/ORBS people think.
The addition of DULs is probably because that open relay stuff isn't such a big deal anymore. So they have to find more reasons to exist.
"They're supposed to complain...". Not bothered who complains. The receipients are just expected to lose business. If they don't notice they're losing business then good for them.
Their competitors just get more customers if their competitors don't use ISPs which incorrectly use MAPS.
At my office we have an antispam software that uses something like a MAPS list, and it blocks mail from DULs. I checked today and there were a few genuine requests for quotations and other important emails.
Now that really sucks. But I think it isn't a hard block and we can probably adjust the weights to not count the IP lists as so important. Or we may just switch to a different software.
The trouble is with ISPs who just 100% block messages from IPs on such lists.
Nah MAPs and other RBLs just suck. They don't really help reduce spam for the amount of false positives they generate.
But if anybody really wants to use the black lists they should use them as _part_ of the antispam process. If a sending IP is on the list, you ALLOW the connection, but you weight the message on the more spammy side. That way if the message is not spam it still has a chance to vindicate itself.
The popular method of 100% blocks is plain stupid and outdated.
"Why people don't switch to djdns (which install in stupid places, is mostly unmaintained, is written by an offensive asshole, and that you cannot fork/modify) ?"
Unmaintained? It ain't broke, so there's no need to fix it every year, unlike plenty of other software.
Sure, there are plenty of people who find DJB offensive. But you'd think that because of that there'd be plenty of people hoping to rub his face in the dirt by finding security holes in his software. So far I don't even see anybody finding security bugs in djbdns. Please do point them out. He has a USD500 guarantee on his software. While the awarding of the USD500 bucks could be disputed, you'd be pretty famous in the IT security field if you do find a valid one.
In contrast finding a security bug in ISC's software is just going to give you like a week or two of fame.
Can't fork/modify? I'm not going to bother figuring out what copyright laws apply in your country wherever it is.
Imagine if a huge boulder fell on you. You'd be spending the next few centuries digging your way out. That is if you don't get confused and dig the wrong way.
Or some sadist found out you can't die but you aren't so strong and still feel pain.
If you can't die, but remain vulnerable, eternity is a very long time. Eventually something bad will happen to you.
If they aren't happy with the way things work most people will just use something else. They're not fixers.
Two "clipboards" is too much for most people to handle. Add the fact that some apps want you to use alt-c instead of ctrl-c, some support ctrl-insert in addition to whatever they show in the edit menu, some need ctrl-shift-c, and some only work with the middle mouse button crap - you're looking at a big mess.
If things don't improve most people _should_ continue to use windows OR MacOSX, and leave the complex stuff to the clever people.
Most people are creatures of habit. They like routine. Stuff must behave the same way. Heck if their routine is backing stuff up and reformating and reinstalling windows, so be it, you'd be surprised how many people prefer living in a "predictable" world no matter how tedious it is. They get annoyed with windows not because it crashes, but because it often crashes "unpredictably". If it only crashes if they do XYZ, then they just won't do XYZ. They're not the sort who'd go - ah it crashes, let's see if it's exploitable.
How does Apple do it with MacOSX and various X apps? They're the one who popularized the copy and paste thing with their first Mac OS. Windows just copied the stuff. AFAIK the Mac and Windows style is very similar. So it's better if the X windows people follow the same concept even if it means breaking a few eggs along the way - I personally think breaking the old X apps for better Linux desktop adoption is worth it. Most of the old X apps aren't worth running. Heck you want the old behaviour make it an option and not the default. The smart ones can figure it out.
The stupid ones are more plentiful than the smart ones, so if people _really_ want adoption by the _masses_, then cater for the stupid. So what if the smart ones get annoyed. They're smart, they'll learn, they can workaround almost anything. Do it right and they might even think it's an interesting challenge.
Those who say there isn't a problem, are actually people who don't get it. We're talking about usability. The fact that many people think there is a problem, means there is one. Even if it's because they are confused over something, it means something is confusing and thus there is a problem.
If the OSS developers are happy with the way things work, I'm sure MS will be very happy too and some of the smart ones aren't so smart after all.
Maybe you just knew how to use those muscles better. Boys/Guys tend to have more fighting and wrestling experience - even theoretical knowledge from watching others helps.
"They break up words with html tags, use all sorts of MIME hackery etc, to trick the filter into not using the right set of data"
Right, and grandma gets legitimate mail that does this all the time... That spam technique is short sighted.
An antispam filter worth anything would start regarding a message as more spammy when it detects something like this.
I don't write spam or antispam stuff but I bet a hard spam technique to catch would be those with a line of text (click here for more info) a link and optionally followed by a random joke or random legit text from the internet/wikipedia/some joke site/darwin award etc.
Go ahead I bet you can think of even more devious methods. I'm sure tons of friends/relatives/colleagues/sites have sent you legit stuff that follows this style.
Naive bayesian filters could start getting more false positives if users started telling them that this sort of stuff is spam. I daresay it would be better to categorize this as "unsure"/"indeterminable".
You'd need whitelists and probably blacklists (known spammer websites/urls) to stop this. But then they'd use tinyurl and similar stuff.
Heck they could use I'm feeling lucky on google with the right keywords. Fortunately I'm not a spammer.
Disclaimer: The company I work for resells stuff like puremessage (and provides other IT securitystuff/services).
Just wait till the spammers start training their spam generators through dspam.
Furthermore: some people/sites just write messages that look like spam.
A legit sender could say this: "Activate your registration now - click here [url link]". I mean what else do you want them to say without wasting bandwidth?
A spammer could send nearly the same message.
So you'd probably have to blacklist/whitelist the urls they link to.
I think having decoy email accounts to identify spam could be a useful tool.
Keep your wits about you, stop listening to music and being oblivious in dangerous environments. If it's not dangerous then yeah go on listening to music, nothing to worry about.
Why would muggers pick on difficult targets? They'd go for easy targets.
If it's just people looking for a way to get their next drug high, you may be able to negotiate a deal - they get money for their hit, you get to keep most of your stuff and no blood spilled. They may not even remember you - it's nothing personal - all they care about is getting their next high.
Otherwise, if it looks real bad, take out wallet, take the cash out, drop the cash on ground and run. If he still goes for you, chuck some stuff, if he doesn't stop to collect the stuff it may be personal (or mistaken identity), good luck...
Well not much comfort if he shoots you, even if you get him after it still sucks.
In my country fortunately in general the armed robbers usually have one or two crappy gun amongst them, and the police are the ones with decent firearms. Better than the other way round where you have crooks with SMGs and the police with cheap revolvers. Armed robbery is rare. Snatch thefts from motocycles are more common - whiz past and yank the stuff off. A victim died when she fell and hit her head. But I still think it's safer than with tons of guns around.
But trying to reduce gun ownership in the US is like trying to get the countries with nukes to disarm. It's probably too late. That said the increase in deadly weapons doesn't make things safer despite gun ownership proponents - it's just an arms race.
Coz the deadlier the weapons available, the greater the radius of vulnerability. While you can more easily defend yourself, the attackers will have similar weapons too. It's like going back to living in a jungle with lots of predators, rather than a civilized city. With the large predators going extinct the jungles are getting pretty safe;).
Back to the story, if I'm a mugger I'd go for those who aren't alert - so those oblivious people listening to their ipods will make ripe targets. Usually silly to go for someone who is obviously alert even if they're only 75% of your capabilities in a fight. Predators that keep going for prey that are ready and able to fight back don't tend to do very well.
LISP is like building stuff at a molecular level. You can build anything. Heck you can build stuff to build stuff. Many people even build stuff to build stuff to build stuff and so on till end of recursion.
But since I'm a crappy programmer I use Perl + CPAN - which is like building stuff using prefab/off the shelf components. Sure the results aren't "perfect" down to a molecular level and the CS ppl will probably puke, but hey it works, and I'll cover reasons that'll relate to why LISP isn't so popular.
From article: "Finally, consider your favorite debugger---or rather, compare it to your favorite editor. You can probably write macros for the latter; why can't you for the former? And why can't programmers include code in libraries to control how debuggers display the data structures those libraries create?"
"programmers will be able to extend the syntax of programming languages"
Maybe that's fine if there were only one programmer in the world.
If you computer language can _easily_ call and talk to programs written in other languages, then in a way it is extensible to different compilers with different syntaxes etc.
Sure it's not any arbitrary compiler or syntax BUT is that really a problem? If the compiler and syntax ALREADY exists, other people know what you are doing, and there's already lots of code you can reuse - you don't have to reinvent the wheel.
There may be 1000 different _good_ reasons why the existing compiler and syntax is that particular way, and you in your limited experience don't know yet. Similarly there may be 1000 different good reasons why a prefab module is a particular way, so it's often better not to make your own (barring restrictions like copyright/patents etc).
Most programmers don't really write anything really new that requires a totally unique syntax/compiler.
So is it really better if it's easy for programmers to reinvent stuff? They'd most likely screw up.
Sometimes to make things easier overall, it is best to make some things hard, so that things that shouldn't be done, rarely or never get done.
In theory it would be nice to be able to specify the exact colour you want for the paint factory to make. But often it is quicker and better to just pick an already prepared paint in a tin. Plus you can always mix colours, so what if it still isn't 100% exactly the colour you want, it's close enough for most people.
Just make it easier to mix colours.
BTW: One man's impedance mismatch is another man's layer of abstraction/API.
Might wish to spend some of it on better voting systems.
Funny how the US of A is willing to spend billions choosing the Gov in Iraq, but have to resort to proven crap like Diebold's systems to choose the US government.
But maybe there really isn't a choice and it's all going to be satisfactorily rigged...
There are plenty of smart people who have proposed pretty decent systems for anonymous auditable voting. But no, Diebold and Friends are the choice.
Slashdot Mirror
He has his reasons, and extremely often he is right. (I think he's wrong about that proposal about the next gen email method tho, not that it matters).
:)
Yeah he is arrogant and often obnoxious but I actually regard this as a feature in this case.
Why? He must have pissed off thousands of people by now (many of them fairly smart too). The way these people can get back at him where it hurts him is to find a security flaw in his program and rub his nose in it.
The fact that I haven't seen any qmail and djbdns exploits (ala remote root or remote user ) despite all that is gives me more confidence in the security of his software.
Hmm, some RSI-like symptoms are actually due to nerve problems. Maybe in this case the nerves are weakened and so are more sensitive to just a bit of squishing.
How did you know they're spam?
Did you actually test MAPS/ORBS to see how accurate they are? Is your mailserver configured to do further processing on mails from IPs listed on MAPS/ORBS?
At home I regularly get 100-200+ spams a day. Spambayes screens most of them with minimal false positives (there were a few false positives during the initial training period, zero since then - I'm still testing it ). Even my personal naive keyword filtering screened out plenty of spam with very minimal false positives.
Disclaimer: My company resells Puremessage - it can be configured to take into consideration a listing on DNS blacklists. Taking into consideration means adding spamminess points, but still going through with the rest of the spamminess and hamminess tests.
Personally the MAPS/ORBS stuff is pretty shoddy. In the old days they even let you download lists of open relays. Now that must have been pretty useful for spammers and some other folk... To me that indicated to me how the MAPS/ORBS people think.
The addition of DULs is probably because that open relay stuff isn't such a big deal anymore. So they have to find more reasons to exist.
"They're supposed to complain...". Not bothered who complains. The receipients are just expected to lose business. If they don't notice they're losing business then good for them.
Their competitors just get more customers if their competitors don't use ISPs which incorrectly use MAPS.
At my office we have an antispam software that uses something like a MAPS list, and it blocks mail from DULs. I checked today and there were a few genuine requests for quotations and other important emails.
Now that really sucks. But I think it isn't a hard block and we can probably adjust the weights to not count the IP lists as so important. Or we may just switch to a different software.
The trouble is with ISPs who just 100% block messages from IPs on such lists.
Nah MAPs and other RBLs just suck. They don't really help reduce spam for the amount of false positives they generate.
But if anybody really wants to use the black lists they should use them as _part_ of the antispam process. If a sending IP is on the list, you ALLOW the connection, but you weight the message on the more spammy side. That way if the message is not spam it still has a chance to vindicate itself.
The popular method of 100% blocks is plain stupid and outdated.
A lot of times the ISP's SMTP server is configured to reject mails with return paths that don't match any of the ISP's served domains.
And if your domain isn't the ISP's domain, you get a 5xx error when you do the MAIL FROM:youremailaddr
If you change that email addr, you don't get the bounces.
If you change your email address to one of the ISPs then you don't get to use your domain for email.
"Why people don't switch to djdns (which install in stupid places, is mostly unmaintained, is written by an offensive asshole, and that you cannot fork/modify) ?"
Unmaintained? It ain't broke, so there's no need to fix it every year, unlike plenty of other software.
Sure, there are plenty of people who find DJB offensive. But you'd think that because of that there'd be plenty of people hoping to rub his face in the dirt by finding security holes in his software. So far I don't even see anybody finding security bugs in djbdns. Please do point them out. He has a USD500 guarantee on his software. While the awarding of the USD500 bucks could be disputed, you'd be pretty famous in the IT security field if you do find a valid one.
In contrast finding a security bug in ISC's software is just going to give you like a week or two of fame.
Can't fork/modify? I'm not going to bother figuring out what copyright laws apply in your country wherever it is.
Well if you force someone to learn do you increase their potential choices?
:).
They still die in the end, so what's the time period for the number of choices?
Maybe it's a lesser crime to steal everything from someone who's definitely going to die soon
"but as we know we are overpopulated ...."
That's too easily changed. And will become easier and easier. It is easier to destroy than to sustain.
Imagine if everyone had immense power and all had access to a "Kill Everybody" Button.
Plus all it takes is a big asteroid/comet or two and we'd be a lot less populated.
Highlander? Being immortal like that is a curse.
Imagine if a huge boulder fell on you. You'd be spending the next few centuries digging your way out. That is if you don't get confused and dig the wrong way.
Or some sadist found out you can't die but you aren't so strong and still feel pain.
If you can't die, but remain vulnerable, eternity is a very long time. Eventually something bad will happen to you.
And you will wish you were dead.
If they aren't happy with the way things work most people will just use something else. They're not fixers.
Two "clipboards" is too much for most people to handle. Add the fact that some apps want you to use alt-c instead of ctrl-c, some support ctrl-insert in addition to whatever they show in the edit menu, some need ctrl-shift-c, and some only work with the middle mouse button crap - you're looking at a big mess.
If things don't improve most people _should_ continue to use windows OR MacOSX, and leave the complex stuff to the clever people.
Most people are creatures of habit. They like routine. Stuff must behave the same way. Heck if their routine is backing stuff up and reformating and reinstalling windows, so be it, you'd be surprised how many people prefer living in a "predictable" world no matter how tedious it is. They get annoyed with windows not because it crashes, but because it often crashes "unpredictably". If it only crashes if they do XYZ, then they just won't do XYZ. They're not the sort who'd go - ah it crashes, let's see if it's exploitable.
How does Apple do it with MacOSX and various X apps? They're the one who popularized the copy and paste thing with their first Mac OS. Windows just copied the stuff. AFAIK the Mac and Windows style is very similar. So it's better if the X windows people follow the same concept even if it means breaking a few eggs along the way - I personally think breaking the old X apps for better Linux desktop adoption is worth it. Most of the old X apps aren't worth running. Heck you want the old behaviour make it an option and not the default. The smart ones can figure it out.
The stupid ones are more plentiful than the smart ones, so if people _really_ want adoption by the _masses_, then cater for the stupid. So what if the smart ones get annoyed. They're smart, they'll learn, they can workaround almost anything. Do it right and they might even think it's an interesting challenge.
Those who say there isn't a problem, are actually people who don't get it. We're talking about usability. The fact that many people think there is a problem, means there is one. Even if it's because they are confused over something, it means something is confusing and thus there is a problem.
If the OSS developers are happy with the way things work, I'm sure MS will be very happy too and some of the smart ones aren't so smart after all.
Yeah. I'm still puzzled why some people prefer to be called disabled instead of handicapped.
Prob is you might hit some 2GB/4GB file size limits.
Maybe you just knew how to use those muscles better. Boys/Guys tend to have more fighting and wrestling experience - even theoretical knowledge from watching others helps.
How'd your CPU die anyway? Curious.
Seriously. No puns please.
The other advantage is the licensing. With Postgresql it's pretty clear what you can or cannot do. Esp as a system integrator or application dev.
"They break up words with html tags, use all sorts of MIME hackery etc, to trick the filter into not using the right set of data"
Right, and grandma gets legitimate mail that does this all the time... That spam technique is short sighted.
An antispam filter worth anything would start regarding a message as more spammy when it detects something like this.
I don't write spam or antispam stuff but I bet a hard spam technique to catch would be those with a line of text (click here for more info) a link and optionally followed by a random joke or random legit text from the internet/wikipedia/some joke site/darwin award etc.
Go ahead I bet you can think of even more devious methods. I'm sure tons of friends/relatives/colleagues/sites have sent you legit stuff that follows this style.
Naive bayesian filters could start getting more false positives if users started telling them that this sort of stuff is spam. I daresay it would be better to categorize this as "unsure"/"indeterminable".
You'd need whitelists and probably blacklists (known spammer websites/urls) to stop this. But then they'd use tinyurl and similar stuff.
Heck they could use I'm feeling lucky on google with the right keywords. Fortunately I'm not a spammer.
Disclaimer: The company I work for resells stuff like puremessage (and provides other IT securitystuff/services).
Just wait till the spammers start training their spam generators through dspam.
Furthermore: some people/sites just write messages that look like spam.
A legit sender could say this: "Activate your registration now - click here [url link]". I mean what else do you want them to say without wasting bandwidth?
A spammer could send nearly the same message.
So you'd probably have to blacklist/whitelist the urls they link to.
I think having decoy email accounts to identify spam could be a useful tool.
Keep your wits about you, stop listening to music and being oblivious in dangerous environments. If it's not dangerous then yeah go on listening to music, nothing to worry about.
Why would muggers pick on difficult targets? They'd go for easy targets.
If it's just people looking for a way to get their next drug high, you may be able to negotiate a deal - they get money for their hit, you get to keep most of your stuff and no blood spilled. They may not even remember you - it's nothing personal - all they care about is getting their next high.
Otherwise, if it looks real bad, take out wallet, take the cash out, drop the cash on ground and run. If he still goes for you, chuck some stuff, if he doesn't stop to collect the stuff it may be personal (or mistaken identity), good luck...
Well not much comfort if he shoots you, even if you get him after it still sucks.
;).
In my country fortunately in general the armed robbers usually have one or two crappy gun amongst them, and the police are the ones with decent firearms. Better than the other way round where you have crooks with SMGs and the police with cheap revolvers. Armed robbery is rare. Snatch thefts from motocycles are more common - whiz past and yank the stuff off. A victim died when she fell and hit her head. But I still think it's safer than with tons of guns around.
But trying to reduce gun ownership in the US is like trying to get the countries with nukes to disarm. It's probably too late. That said the increase in deadly weapons doesn't make things safer despite gun ownership proponents - it's just an arms race.
Coz the deadlier the weapons available, the greater the radius of vulnerability. While you can more easily defend yourself, the attackers will have similar weapons too. It's like going back to living in a jungle with lots of predators, rather than a civilized city. With the large predators going extinct the jungles are getting pretty safe
Back to the story, if I'm a mugger I'd go for those who aren't alert - so those oblivious people listening to their ipods will make ripe targets. Usually silly to go for someone who is obviously alert even if they're only 75% of your capabilities in a fight. Predators that keep going for prey that are ready and able to fight back don't tend to do very well.
Unfortunately there are too many like him.
Hopefully you can find a place to live where guns and ammo are hard to find to keep yourself safe from trigger-happy people.
"When was the last time Hillary Duff was in a shootout with the police?"
Duffy the Donut Cop Slayer. Nah. Doesn't work.
LISP is like building stuff at a molecular level. You can build anything. Heck you can build stuff to build stuff. Many people even build stuff to build stuff to build stuff and so on till end of recursion.
But since I'm a crappy programmer I use Perl + CPAN - which is like building stuff using prefab/off the shelf components. Sure the results aren't "perfect" down to a molecular level and the CS ppl will probably puke, but hey it works, and I'll cover reasons that'll relate to why LISP isn't so popular.
From article: "Finally, consider your favorite debugger---or rather, compare it to your favorite editor. You can probably write macros for the latter; why can't you for the former? And why can't programmers include code in libraries to control how debuggers display the data structures those libraries create?"
"programmers will be able to extend the syntax of programming languages"
Maybe that's fine if there were only one programmer in the world.
If you computer language can _easily_ call and talk to programs written in other languages, then in a way it is extensible to different compilers with different syntaxes etc.
Sure it's not any arbitrary compiler or syntax BUT is that really a problem? If the compiler and syntax ALREADY exists, other people know what you are doing, and there's already lots of code you can reuse - you don't have to reinvent the wheel.
There may be 1000 different _good_ reasons why the existing compiler and syntax is that particular way, and you in your limited experience don't know yet. Similarly there may be 1000 different good reasons why a prefab module is a particular way, so it's often better not to make your own (barring restrictions like copyright/patents etc).
Most programmers don't really write anything really new that requires a totally unique syntax/compiler.
So is it really better if it's easy for programmers to reinvent stuff? They'd most likely screw up.
Sometimes to make things easier overall, it is best to make some things hard, so that things that shouldn't be done, rarely or never get done.
In theory it would be nice to be able to specify the exact colour you want for the paint factory to make. But often it is quicker and better to just pick an already prepared paint in a tin. Plus you can always mix colours, so what if it still isn't 100% exactly the colour you want, it's close enough for most people.
Just make it easier to mix colours.
BTW: One man's impedance mismatch is another man's layer of abstraction/API.
Must be an alternate universe for 5th digit of Pi to be 9. Unless you're talking about something else.
Might wish to spend some of it on better voting systems.
Funny how the US of A is willing to spend billions choosing the Gov in Iraq, but have to resort to proven crap like Diebold's systems to choose the US government.
But maybe there really isn't a choice and it's all going to be satisfactorily rigged...
There are plenty of smart people who have proposed pretty decent systems for anonymous auditable voting. But no, Diebold and Friends are the choice.