I mean, they've got one division dedicated to optimizing the throughput of their hardware to deliver the maximum spam/sec
I'm confused -- should there be a ban on efficient mail gateways or something? Spammers don't have the budget for Ironport boxen. They especially won't have the budget to pay for an Ironport bond, because every single complaint will cost them money.
You do realize of course that two of the checks in SpamAssassin are for Spamcop and the Ironport Bonded Sender whitelist?
Now am I that far off? Are there redeeming qualities about SpamCop that I'm overlooking that make this blocklist a good thing? Who controls who's blocked? Is it fair? Is there a human contact when things go awry?
Spamcop blocks nothing. It simply tells you whether an IP address is on the list and provides you with a URL to use to make the final determination yourself. Spamcop itself does not recommend automatically blocking based on SC listing alone. Spamcop is quite trigger happy because it's largely automated (and unfortunately the DNSBL doesn't include things like volume statistics to help automate it -- I suspect that'd be a commercial value-add). If you want a more carefully researched blacklist, I recommend the spamhaus SBL.
The contact process is documented on spamcop's site. Humans do read and act on it. They're generally a lot more reasonable than, say, the folks on usenet you'll have to speak to if you wind up on SPEWS.
I'd be astroturfing if I told you the overall "safe" solution I recommend, so I'll leave it there.
I wasn't able to completely figure out what the Stackless project is trying to achieve. Prolly since I'm not really a programmer. Can someone shed some light on this?
Stackless, which incidentally isn't the best name for it (it's based on an architecture that is best described now as "many stacks", not stackless), is a patch to Python that started off as offering first-class continuations (a continuation can be described as taking a snapshot of the run state and providing a sort of "goto" to get back to that snapshot), but it quickly moved away from that into a concurrency model called microthreads. Microthreads in a nutshell are a way of structuring programs as if they were multithreaded, but unlike "real" threads, only one thread ever runs at once, so either a truly concurrent scheduler must pre-empt them, or a microthread must yield to allow another to run.
The advantage to microthreads is mostly about syntax -- you can write programs as if they were multithreaded, but without having to lock anything, or use awkward bottleneck constructs like thread pools. With added features like channels (a pipe between microthreads), the fact that switching between microthreads is faster than calling a function, and the upcoming feature of 3.0 that lets you pickle microthreads, Stackless really should deserve a spot in Python core (it even dropped continuations, which was Guido's main objection).
Back to the topic: Ironport appliances run on Stackless Python, using many thousands of microthreads (presumably one for every mail that comes in). Another production use of Stackless is the EVE Online MMORPG. Ironport must keep Tismer busy on other things tho -- Stackless hasn't seen a public update in many months...
> I find it very strange that Boies has reached this superstar status on the back of what is by any measure a lacklustre litigation record.
Maybe it's because we're all mesmerized by his eclectic blend of haunting lyrics, thumping funk, and infectious grooves... Oh wait, David Boies? Nevermind...
On an unrelated note, I'm glad to see IBM handling the case the way they are. I'm sincerely hoping that IBM is subpoening Canopus in order to pierce the corporate veil.
1) Canopy, not Canopus.
2) IBM's subpoenas have nothing to do with "piercing the corporate veil". In fact it makes it even more of an action against a corporation than any real persons.
3) You don't push to have lawyers disbarred no matter what meanie stinkie poos you think their clients are. Boies is just another shark in the pool, and quite probably as professional as they come -- they're not the ones issuing the press releases.
for i in `cat/etc/ld.so.conf`/lib/usr/lib;do ls $i/ libraryName.so*;done 2>/dev/null | grep -v @ | sed 's/.*\.so\.//g' | sed 's/\*//g'
I suppose that was actually an attempt to be funny... Anyway, this sort of thing is precisely what ports does, and it front-ends a real package manager at that (it builds and installs binary packages). I've always found ports a breath of fresh air compared to RPM. I always build apache manually because there's just so many different options, and often I have a couple different httpds installed. With RPM, neither of the apaches I built are even considered to be visible. With debian, I have to go through some bureaucratic "fake package build" (debianites sure do love procedure). With ports, I just make install, and it gets found automatically.
CommandPrompt Mammoth PostgreSQL for Win32, Mac OS X, Linux http://www.commandprompt.com/entry.lxp?lxpe =295
Commercial only, even the commercial version is crippled. Incompatible with the very existence of cygwin or mysql on the machine.
dbExperts PostgreSQL for Windows, Mac OS X, Linux http://www.dbexperts.net/postgresql
Commercial only.
PowerGres (threaded Windows PostgreSQL http://osb.sra.co.jp/PowerGres/introdu ction-en.php
Commercial only, japanese only (mostly).
Re:Lots of people don't make the connection but
on
SCO News Roundup
·
· Score: 1
This is quite an insightful observation (one of the few areas where the "insightful" moderation flag is apropos), but I wouldn't worry about the vast might of Microsoft coming down to bury the industry in paperwork. Because to become an IP company that does nothing but pursue legal action, a company generally has to become a miserable abject utter failure like SCO. If MS gets to that point, they'll have spun off, gotten bought off in chunks, and generally deteriorated to the point where they'll present as much a threat to the whole industry as the measly pissants of SCO.
No, there wasn't much of a point to this reply except to give my incisive and witty asides on SCO, surely an iconoclastic viewpoint I know... Come to think of it this characterizes a lot of/. articles on SCO, and unfortunately, Groklaw. Is there any forum like what groklaw was that isn't currently colored by rah-rah boosterism?
> Well, in the short run, loan referrals are STILL worth $50, so spamming a spammer who is doing that will result in an insane windfall for said spammer
Until the spammer is prosecuted for criminal fraud for knowingly giving bad referrals.
My main problem with the "flood 'em with bogus info" approach is that for product spams, it might just take a merchant tool a couple minutes to weed them all out unless you used honest to goodness working credit card numbers (then you're either giving away all your money or committing wire fraud). Mortgage referral spams are probably one of the only spams this trick would work with. That and nigerian spams, though that's even easier since it can be done entirely through email.
I'm as much against plagarism as the next guy, perhaps more so, but can an Anonymous Coward really be said to have plagarized anything? It's not as if Mr AC can take credit for it...
This just kills me, it's such a no-brainer to defeat. Have a security setting in IE on URL's with an '@' in them. Use standard zone security policies of "allow/deny/ask", obviously using the site to the right of the @ as the principal for zone determination. Microsoft could issue this as a simple patch, and set it to "warn" by default (half the people will click "never show me this again" just to dismiss MS's incessant modal popups, but at least they were fairly warned). With some extra thought (perhaps falsely assuming the prerequisite of basic thought is met by Microsoft), one could include extra security and real nasty warnings using sophisticated textual analysis on the user/pass combo, i.e. it contains a dot.
I worked on a mud in college that allowed user specific code. The head programmer had written a light VM that ran most of the actual mud code. It sandboxed the user land stuff and your code could do whatever you wanted within the context of the syntax and you weren't bringing the MUD down. However, the VM was very memory hungry and really needed to be re-written by someone who knows VM's. He made it all up as he went.
Sounds a lot like MOO... MOO is really quite nice, but it's definitely an ageing creature and needs to be brought up to date. LambdaMOO has an RPG with some pretty sophisticated rule mechanics thanks to the nature of MOO, but being that text MUDs in general are in such a steep decline, it's pretty much barren. That and LambdaRPG is more of a novelty within an otherwise completely social-oriented (talk) MUD, so it doesn't quite get the focus it would deserve from dedicated gamers.
I found that Python lent itself quite nicely to stack-level security like MOO's, and I was able to create a multi-user Python with a couple days worth of hacking -- in fact just locking down the configuration to pre-set the ACL's on some builtins took longer than implementing the mechanics of those ACL's in the first place. It was more a curiosity though -- retrofitting an existing VM to run mutually hostile code is not going to provide the best security in the world.
Longhorn will feature a task-based (or "iterative") interface that goes far beyond the task-based interface found today in Windows XP
Gee I should freakin HOPE it's better than the so-called "task based" interface in XP. Putting the right-click actions into a sidebar does not a new workflow make. Where are the "add new task" buttons? Where's the systemwide macro editor that can record actions into VB code? Hey, it was good enough for Office (oh wait, it can't produce VB code either).
Longhorn will of course promise nothing at all specific, just a "richer user experience". Ballocks I say. Not that I expect to see anything like this in any *nix gui either...
That's one of the things I like best about Gentoo's root shell installer. It immediately gets rid of people that are intimidated by that sort of thing, and prevents them from sucking up tons of attention on mailing lists or forums. The difficulty of the installer should be like those little signs in front of rides at amusement parks: "You must be this tall to ride."
This blatant trolling was modded up +5 insightful? Ook ook, grog, has lots of words, man talk smart, he smart, me mod up.
Election commissioners are usually appointed, not elected (funny when you think about it), so I suspect that's going to be one of the last things Lamone says. Rubin's message was well-received, and at least one politician (a republican at that) publically expressed disappointment at the non-response of the election commission.
Summary: Short Diebold, they're going to lose a lot of contracts.
Nokia does not have an issue with exploding batteries. All cell phones have issues with exploding cheap-ass third party batteries. Lithium Ion batteries are volatile suckers if you don't have the proper circuitry in them.
Nokia has issues with designing a crappy handheld game system for their first try, then when they get called on it, rather than admit honest failure and trying again, they act like a cornered animal. Too bad, I really like the phones (mostly because they have no antenna to snag or flipout to break off), but I'm not sure how committed they'll be to making those continue to work well now -- they might just decide their customers whine too much to be worth listening to.
The "IANAL" should be as unnecessary as reprinting your name in your post. If one is a lawyer, they'll say so, then their boilerplate should read:
"IAAL, but I am probably not licensed in your state. [insert opinion]. None of this is to be taken as legal advice -- please contact a licensed attourney in your state if you want legal advice."
An oxymoron would be something like: A version of Windows shipped with Windows Media Player is a superior operating system.
No that's simply a statement which you believe is false. An oxymoron is an adjective or verb that is contradictory with the attributes of its noun, e.g. "Microsoft Works".
> open source != free... you can definately charge for open-source.
For most open source, you can't stop the first person who pays for it to turn that equality into an equality (and yes, I'm still talking free as in beer). Otherwise you're talking about source licensing of otherwise closed source, and frankly if it's not open to the public, it's not particularly open source.
Bah. I'm getting suckered into one of these aforementioned displays already.
Please to partially evaluate your function pointer then stuff it into a queue of callbacks that counts the number of times this callback has been queued, without actually scanning the queue, because it may be too large. In C.
Some nitwit moderator called the last post a troll. So mod this repost too:
Why do
template
<typename A, typename B>
A f(B b)
{ ...
}
when you can do
let f b =...
Seriously, go and read Modern C++ design, because that's the least of the syntactic salt of C++ templates. It may be a powerful feature, but it's so encrusted with syntactic baggage that it often obscures the actual logic of the program. For the absurd conclusion to the horrors of template syntax, check out the Loki library by the author of the aforementioned book. This library, recall, is supposed to make powerful constructs easier to work with...
Incidentally, to the idiot moderator, the preceding example was ocaml, not BASIC.
Slashdot Mirror
I mean, they've got one division dedicated to optimizing the throughput of their hardware to deliver the maximum spam/sec
I'm confused -- should there be a ban on efficient mail gateways or something? Spammers don't have the budget for Ironport boxen. They especially won't have the budget to pay for an Ironport bond, because every single complaint will cost them money.
I personally opt for SpamAssassin Milter
You do realize of course that two of the checks in SpamAssassin are for Spamcop and the Ironport Bonded Sender whitelist?
Now am I that far off? Are there redeeming qualities about SpamCop that I'm overlooking that make this blocklist a good thing? Who controls who's blocked? Is it fair? Is there a human contact when things go awry?
Spamcop blocks nothing. It simply tells you whether an IP address is on the list and provides you with a URL to use to make the final determination yourself. Spamcop itself does not recommend automatically blocking based on SC listing alone. Spamcop is quite trigger happy because it's largely automated (and unfortunately the DNSBL doesn't include things like volume statistics to help automate it -- I suspect that'd be a commercial value-add). If you want a more carefully researched blacklist, I recommend the spamhaus SBL.
The contact process is documented on spamcop's site. Humans do read and act on it. They're generally a lot more reasonable than, say, the folks on usenet you'll have to speak to if you wind up on SPEWS.
I'd be astroturfing if I told you the overall "safe" solution I recommend, so I'll leave it there.
I wasn't able to completely figure out what the Stackless project is trying to achieve. Prolly since I'm not really a programmer. Can someone shed some light on this?
Stackless, which incidentally isn't the best name for it (it's based on an architecture that is best described now as "many stacks", not stackless), is a patch to Python that started off as offering first-class continuations (a continuation can be described as taking a snapshot of the run state and providing a sort of "goto" to get back to that snapshot), but it quickly moved away from that into a concurrency model called microthreads. Microthreads in a nutshell are a way of structuring programs as if they were multithreaded, but unlike "real" threads, only one thread ever runs at once, so either a truly concurrent scheduler must pre-empt them, or a microthread must yield to allow another to run.
The advantage to microthreads is mostly about syntax -- you can write programs as if they were multithreaded, but without having to lock anything, or use awkward bottleneck constructs like thread pools. With added features like channels (a pipe between microthreads), the fact that switching between microthreads is faster than calling a function, and the upcoming feature of 3.0 that lets you pickle microthreads, Stackless really should deserve a spot in Python core (it even dropped continuations, which was Guido's main objection).
Back to the topic: Ironport appliances run on Stackless Python, using many thousands of microthreads (presumably one for every mail that comes in). Another production use of Stackless is the EVE Online MMORPG. Ironport must keep Tismer busy on other things tho -- Stackless hasn't seen a public update in many months...
> I find it very strange that Boies has reached this superstar status on the back of what is by any measure a lacklustre litigation record.
Maybe it's because we're all mesmerized by his eclectic blend of haunting lyrics, thumping funk, and infectious grooves... Oh wait, David Boies? Nevermind...
On an unrelated note, I'm glad to see IBM handling the case the way they are. I'm sincerely hoping that IBM is subpoening Canopus in order to pierce the corporate veil.
1) Canopy, not Canopus.
2) IBM's subpoenas have nothing to do with "piercing the corporate veil". In fact it makes it even more of an action against a corporation than any real persons.
3) You don't push to have lawyers disbarred no matter what meanie stinkie poos you think their clients are. Boies is just another shark in the pool, and quite probably as professional as they come -- they're not the ones issuing the press releases.
"Extreme Bugs Found In Slag Dump"
I've written code that could be described like that after it crashed and took everything else with it...
I suppose that was actually an attempt to be funny... Anyway, this sort of thing is precisely what ports does, and it front-ends a real package manager at that (it builds and installs binary packages). I've always found ports a breath of fresh air compared to RPM. I always build apache manually because there's just so many different options, and often I have a couple different httpds installed. With RPM, neither of the apaches I built are even considered to be visible. With debian, I have to go through some bureaucratic "fake package build" (debianites sure do love procedure). With ports, I just make install, and it gets found automatically.
CommandPrompt Mammoth PostgreSQL for Win32, Mac OS X, Linuxe =295
u ction-en.php
http://www.commandprompt.com/entry.lxp?lxp
Commercial only, even the commercial version is crippled. Incompatible with the very existence of cygwin or mysql on the machine.
dbExperts PostgreSQL for Windows, Mac OS X, Linux
http://www.dbexperts.net/postgresql
Commercial only.
PowerGres (threaded Windows PostgreSQL
http://osb.sra.co.jp/PowerGres/introd
Commercial only, japanese only (mostly).
This is quite an insightful observation (one of the few areas where the "insightful" moderation flag is apropos), but I wouldn't worry about the vast might of Microsoft coming down to bury the industry in paperwork. Because to become an IP company that does nothing but pursue legal action, a company generally has to become a miserable abject utter failure like SCO. If MS gets to that point, they'll have spun off, gotten bought off in chunks, and generally deteriorated to the point where they'll present as much a threat to the whole industry as the measly pissants of SCO.
/. articles on SCO, and unfortunately, Groklaw. Is there any forum like what groklaw was that isn't currently colored by rah-rah boosterism?
No, there wasn't much of a point to this reply except to give my incisive and witty asides on SCO, surely an iconoclastic viewpoint I know... Come to think of it this characterizes a lot of
> "Gloria Mundi" probably gets several calls before somebody at the mortgage company remembers high school Latin or a Roman Catholic upbringing.
"Hi, I'm looking for Ms. Gloria Mundi, is she available?"
"I'm sorry, she passed away."
ba-dum-bump.
> Well, in the short run, loan referrals are STILL worth $50, so spamming a spammer who is doing that will result in an insane windfall for said spammer
Until the spammer is prosecuted for criminal fraud for knowingly giving bad referrals.
My main problem with the "flood 'em with bogus info" approach is that for product spams, it might just take a merchant tool a couple minutes to weed them all out unless you used honest to goodness working credit card numbers (then you're either giving away all your money or committing wire fraud). Mortgage referral spams are probably one of the only spams this trick would work with. That and nigerian spams, though that's even easier since it can be done entirely through email.
I'm as much against plagarism as the next guy, perhaps more so, but can an Anonymous Coward really be said to have plagarized anything? It's not as if Mr AC can take credit for it...
This just kills me, it's such a no-brainer to defeat. Have a security setting in IE on URL's with an '@' in them. Use standard zone security policies of "allow/deny/ask", obviously using the site to the right of the @ as the principal for zone determination. Microsoft could issue this as a simple patch, and set it to "warn" by default (half the people will click "never show me this again" just to dismiss MS's incessant modal popups, but at least they were fairly warned). With some extra thought (perhaps falsely assuming the prerequisite of basic thought is met by Microsoft), one could include extra security and real nasty warnings using sophisticated textual analysis on the user/pass combo, i.e. it contains a dot.
I worked on a mud in college that allowed user specific code. The head programmer had written a light VM that ran most of the actual mud code. It sandboxed the user land stuff and your code could do whatever you wanted within the context of the syntax and you weren't bringing the MUD down. However, the VM was very memory hungry and really needed to be re-written by someone who knows VM's. He made it all up as he went.
Sounds a lot like MOO... MOO is really quite nice, but it's definitely an ageing creature and needs to be brought up to date. LambdaMOO has an RPG with some pretty sophisticated rule mechanics thanks to the nature of MOO, but being that text MUDs in general are in such a steep decline, it's pretty much barren. That and LambdaRPG is more of a novelty within an otherwise completely social-oriented (talk) MUD, so it doesn't quite get the focus it would deserve from dedicated gamers.
I found that Python lent itself quite nicely to stack-level security like MOO's, and I was able to create a multi-user Python with a couple days worth of hacking -- in fact just locking down the configuration to pre-set the ACL's on some builtins took longer than implementing the mechanics of those ACL's in the first place. It was more a curiosity though -- retrofitting an existing VM to run mutually hostile code is not going to provide the best security in the world.
Longhorn will feature a task-based (or "iterative") interface that goes far beyond the task-based interface found today in Windows XP
Gee I should freakin HOPE it's better than the so-called "task based" interface in XP. Putting the right-click actions into a sidebar does not a new workflow make. Where are the "add new task" buttons? Where's the systemwide macro editor that can record actions into VB code? Hey, it was good enough for Office (oh wait, it can't produce VB code either).
Longhorn will of course promise nothing at all specific, just a "richer user experience". Ballocks I say. Not that I expect to see anything like this in any *nix gui either...
That's one of the things I like best about Gentoo's root shell installer. It immediately gets rid of people that are intimidated by that sort of thing, and prevents them from sucking up tons of attention on mailing lists or forums. The difficulty of the installer should be like those little signs in front of rides at amusement parks: "You must be this tall to ride."
This blatant trolling was modded up +5 insightful? Ook ook, grog, has lots of words, man talk smart, he smart, me mod up.
Election commissioners are usually appointed, not elected (funny when you think about it), so I suspect that's going to be one of the last things Lamone says. Rubin's message was well-received, and at least one politician (a republican at that) publically expressed disappointment at the non-response of the election commission.
Summary: Short Diebold, they're going to lose a lot of contracts.
> Then they have the issue of exploding batteries
Nokia does not have an issue with exploding batteries. All cell phones have issues with exploding cheap-ass third party batteries. Lithium Ion batteries are volatile suckers if you don't have the proper circuitry in them.
Nokia has issues with designing a crappy handheld game system for their first try, then when they get called on it, rather than admit honest failure and trying again, they act like a cornered animal. Too bad, I really like the phones (mostly because they have no antenna to snag or flipout to break off), but I'm not sure how committed they'll be to making those continue to work well now -- they might just decide their customers whine too much to be worth listening to.
The "IANAL" should be as unnecessary as reprinting your name in your post. If one is a lawyer, they'll say so, then their boilerplate should read:
"IAAL, but I am probably not licensed in your state. [insert opinion]. None of this is to be taken as legal advice -- please contact a licensed attourney in your state if you want legal advice."
I'm not an expert but I play one on slashdot.
An oxymoron would be something like:
A version of Windows shipped with Windows Media Player is a superior operating system.
No that's simply a statement which you believe is false. An oxymoron is an adjective or verb that is contradictory with the attributes of its noun, e.g. "Microsoft Works".
> open source != free... you can definately charge for open-source.
For most open source, you can't stop the first person who pays for it to turn that equality into an equality (and yes, I'm still talking free as in beer). Otherwise you're talking about source licensing of otherwise closed source, and frankly if it's not open to the public, it's not particularly open source.
Bah. I'm getting suckered into one of these aforementioned displays already.
try explaining the good side of open-source programs to people who have never written a program in their life
"It's free."
(this lasts til some other nearby geek decides to put on a dominance display and quibbles over the connotation of "Free")
> I'm not an expert, but I was under the impression that wine quality improves with more sunlight
Nope. Wine quality improves with drought. They need just enough water to get started, but after that, wineries hate what farmers love: rain.
Please to partially evaluate your function pointer then stuff it into a queue of callbacks that counts the number of times this callback has been queued, without actually scanning the queue, because it may be too large. In C.
This is a real world example, incidentally.
Seriously, go and read Modern C++ design, because that's the least of the syntactic salt of C++ templates. It may be a powerful feature, but it's so encrusted with syntactic baggage that it often obscures the actual logic of the program. For the absurd conclusion to the horrors of template syntax, check out the Loki library by the author of the aforementioned book. This library, recall, is supposed to make powerful constructs easier to work with...
Incidentally, to the idiot moderator, the preceding example was ocaml, not BASIC.