I'll assume you've never learned a second language. If you had, you would know that learning the subtleties of another language can take a good decade or sometimes a lifetime. Considering it was a German writing the English translation, I'd give it an A+. Whether they should have hired professional translators as this is a business, is another topic.
Is the light just right for the photo? No, you assfuck photographer--my left side, my left side! Okay... unbutton the top button? More chest hair? Okay, not put them in a collage for the website. Yeah.
I'm studying Computer Engineering. Sure, I don't need the simplicity of Macs, per se, but why should being highly computer literate and wanting a simple computer be mutually exclusive? I'm dying to get my hands on an iBook with OS X. It is absolutly beautiful and great for doing all my daily work. Just because I'm a Vi wizard and grepmaster doesn't mean I don't like using a glossy word processor. Just because I know how to compile the drivers and hack up a linux system to support my particular camera through a series of scripts doesn't mean that method is better. It's not. Simplicity is good. It took some pretty damn good software engineers to put together this entire system and make it easy. So why isn't that desireable on Linux? Oh it is? They why hasn't it been done yet?
It doesn't matter how wonderful, secure, stable and efficient Linux is--it will never take over the desktop until there are gay little wizards and paper clips talking to you, and both major GUI's can come together and standardize or one of them dies (I'd vote for Gnome biting it). The problem with Linux (really *nix in general) is that there are just too many ways to do something which overwhelms new users. I don't think it is so much just not wanting to learn something new. Also a problem is that most average users are oblivious to MS problems--they just don't hear about them, or if they do they don't know how to patch or just don't care because they think security breaches will never happen to them. Unix types are power users. We want everything customized how we want it to a T. Most users just don't care. If they can get their email--great. Just "point and click."
As for business--I see continued growth. With the addition of things like stateful firewalls and journaling filesystems, more business are going to be installing it in more critical applications.
If you've been reading c.l.python for the past month or so, you've been getting a bit of a preview to his work. He's been putting up little pieces of code which are pretty cool and getting advice from the roundtable.
that he was mapping out the human genome on a few G4's--his supercomputers on a chip--in his basement in his spare time and was scheduled to be finished in about 2003 and start a bio tech company. Oh, then he said that with the raging speed, and blazing performance of his mega-cool, and tantalizingly awesome G4, he already rendered all the animations to his next 284 movies Pixar will release in the next 1024 years last night while he was just "taking a dump." Sure enough! "Hot fscking damn" he said--"I'm getting bored--I think I'll calculate the position of Pluto in 3026--the year a complex simulation on his G4 told him Microsoft would see its demise. You see, I'm going to be cryogenically frozen and revive myself in 3026--that's the target date. That's the plan." Meanwhile, I've left Pixar in good hands and will will brainwash the youth of the planet with the films--laden with subliminal propoganda-- I just rendered and pave the way to my triumph. I will use the genome to create hunter-killer types that will go after Microsoft. When asked if he thought Gates had plans for cryogenic storage as well, only a soft audible grunt--aparantly some veiled explicitive or insult could be heard. He was noticibly angered. He then muttered something about using his G4 to find a new element or something.
Well, already the ladies always notice my sleak visor edge and want to play with it, but imagine when they see me grepping and making--We're talkin' 100 Vestal virgins on a tropical island now
One "feature" that bugs the hell out of me is the automatic conversion of >'s in mail and news replies to vertical gray bars. This wreaks havoc in the Python newsgroup where some session code like
>>> spam = "asdf"
>>> 1 + 2
3
>>>
looks more like this...
||| spam = "asdf"
||| 1 + 2
3
|||
Except with really ugly gray vertical lines. This really needs to be an option to turn off. I haven't been able to find the setting in the options, however.
Another book about Microsoft?!
on
The Forever War
·
· Score: 3, Funny
Phillips charged for conspiracy to place "spikes of death," used to maim and kill loggers, in forrests; California enviornmental groups outraged. Activist quoted as saying "damn it, we thought of it first!"
Brent Lundberg: Oddly enough, Jim, the Chinese seem to have delt a crushing blow to the U.S. and Swiss top downhill contenders with times up to 45 seconds lower.
Jim: But the Chinese can't ski!
meanwhile in Beijing:
Xing Zhang: Muhahahahahahahaha
Guitarists have already rejected technically-superior digital solid state amps going back instead to vacuum tubes because of the warmer sound. They won't go digital this time either.
Let's see...
front door --> driveway: scooter
driveway --> parking garage: car
parking garage --> office building: scooter
front door --> cubicle: scooter
cubicle --> upstairs coffee room. Awww fuck, you mean I have to walk?
Taking a hash or SHA digest is bound to fail. One little character is off and the thing fails. That means customized emails, for example.
I'm working on a histogram based filter. You count the number of time a word occurrs in an email and you create a hash based upon the top twenty words and their occurance rate or so. If the spammer changes dear "bob" to dear "fred", bob and fred only occurr once or twice at most and are deemed insignificant by the algorithm and do not affect the hash. The more words you accept, the more accurate the fingerprint, although 20 or so seems to be accurate enough. Other configurable parts of the algorithm allow you to bump the word increment once if it occurs two times or three times instead of a one-to-one increment. Minute changes, therefore, will not affect the hash.
Furthermore, if an exact match is not made, you can keep compiling all histograms by sorting them into groups of like content, and then generate "master" general fingerprints, which can then be used with weights in a fuzzy algorithm to score a message for spammyness. Combined with a threshold (say 50% spamminess) you can decide whether or not you want to reject it. This again is only used if a direct match is not made. (If an exact fingerprint has not yet made it into the db).
This system still plays nice with distributed methods as you are still using a small hash code. If it were employed on one system, it would be easier to keep more detailed records of each hist; not just a hash.
Maybe one of these days I'll have a decent working prototype in python to share.
The best option would be a word count histogram filter. Then the spammer would have to entirely alter their language or sales pitch, which isn't going to happen. Just like handwriting, it is hard to change unless you make a whosale effort at changing it. They're too lazy, too.
Slashdot Mirror
That you'd only be measuring the amount of pr0n being downloaded by physics students... unless you had your own clean segment.
I'll assume you've never learned a second language. If you had, you would know that learning the subtleties of another language can take a good decade or sometimes a lifetime. Considering it was a German writing the English translation, I'd give it an A+. Whether they should have hired professional translators as this is a business, is another topic.
I can't believe its not Linux.
Is the light just right for the photo? No, you assfuck photographer--my left side, my left side! Okay... unbutton the top button? More chest hair? Okay, not put them in a collage for the website. Yeah.
A shopping button on the KDE desktop!
I'm studying Computer Engineering. Sure, I don't need the simplicity of Macs, per se, but why should being highly computer literate and wanting a simple computer be mutually exclusive? I'm dying to get my hands on an iBook with OS X. It is absolutly beautiful and great for doing all my daily work. Just because I'm a Vi wizard and grepmaster doesn't mean I don't like using a glossy word processor. Just because I know how to compile the drivers and hack up a linux system to support my particular camera through a series of scripts doesn't mean that method is better. It's not. Simplicity is good. It took some pretty damn good software engineers to put together this entire system and make it easy. So why isn't that desireable on Linux? Oh it is? They why hasn't it been done yet?
It's just "Mark"... You don't say Punts, Francs, and Americandollars... Furthermore it is Mark singular or plural.
It doesn't matter how wonderful, secure, stable and efficient Linux is--it will never take over the desktop until there are gay little wizards and paper clips talking to you, and both major GUI's can come together and standardize or one of them dies (I'd vote for Gnome biting it). The problem with Linux (really *nix in general) is that there are just too many ways to do something which overwhelms new users. I don't think it is so much just not wanting to learn something new. Also a problem is that most average users are oblivious to MS problems--they just don't hear about them, or if they do they don't know how to patch or just don't care because they think security breaches will never happen to them. Unix types are power users. We want everything customized how we want it to a T. Most users just don't care. If they can get their email--great. Just "point and click."
As for business--I see continued growth. With the addition of things like stateful firewalls and journaling filesystems, more business are going to be installing it in more critical applications.
There was already a story on this earlier this summer.
and a great page on
space clocks and frequency control technology
If you've been reading c.l.python for the past month or so, you've been getting a bit of a preview to his work. He's been putting up little pieces of code which are pretty cool and getting advice from the roundtable.
that he was mapping out the human genome on a few G4's--his supercomputers on a chip--in his basement in his spare time and was scheduled to be finished in about 2003 and start a bio tech company. Oh, then he said that with the raging speed, and blazing performance of his mega-cool, and tantalizingly awesome G4, he already rendered all the animations to his next 284 movies Pixar will release in the next 1024 years last night while he was just "taking a dump." Sure enough! "Hot fscking damn" he said--"I'm getting bored--I think I'll calculate the position of Pluto in 3026--the year a complex simulation on his G4 told him Microsoft would see its demise. You see, I'm going to be cryogenically frozen and revive myself in 3026--that's the target date. That's the plan." Meanwhile, I've left Pixar in good hands and will will brainwash the youth of the planet with the films--laden with subliminal propoganda-- I just rendered and pave the way to my triumph. I will use the genome to create hunter-killer types that will go after Microsoft. When asked if he thought Gates had plans for cryogenic storage as well, only a soft audible grunt--aparantly some veiled explicitive or insult could be heard. He was noticibly angered. He then muttered something about using his G4 to find a new element or something.
Well, already the ladies always notice my sleak visor edge and want to play with it, but imagine when they see me grepping and making--We're talkin' 100 Vestal virgins on a tropical island now
One "feature" that bugs the hell out of me is the automatic conversion of >'s in mail and news replies to vertical gray bars. This wreaks havoc in the Python newsgroup where some session code like
>>> spam = "asdf"
>>> 1 + 2
3
>>>
looks more like this...
||| spam = "asdf"
||| 1 + 2
3
|||
Except with really ugly gray vertical lines. This really needs to be an option to turn off. I haven't been able to find the setting in the options, however.
Oh, wait... nevermind.
Being able to see all those lusty ladies in 3d, it might use too much bandwidth downloading pr0n.
That PayPal horror stories website is running ads for a PayPal competitor -- c2it.com
So this website has a good deal of interest in shocking and horrifying its visitors.
Phillips charged for conspiracy to place "spikes of death," used to maim and kill loggers, in forrests; California enviornmental groups outraged. Activist quoted as saying "damn it, we thought of it first!"
an unemployment webring for the displaced admins and programmers. Oh, wait! Doh!
You idiot! We're trying to sell processors here!
considering I've received 20 virus-laden emails through my @home account in a matter of days.
Brent Lundberg: Oddly enough, Jim, the Chinese seem to have delt a crushing blow to the U.S. and Swiss top downhill contenders with times up to 45 seconds lower.
Jim: But the Chinese can't ski!
meanwhile in Beijing:
Xing Zhang: Muhahahahahahahaha
Guitarists have already rejected technically-superior digital solid state amps going back instead to vacuum tubes because of the warmer sound. They won't go digital this time either.
Let's see...
front door --> driveway: scooter
driveway --> parking garage: car
parking garage --> office building: scooter
front door --> cubicle: scooter
cubicle --> upstairs coffee room. Awww fuck, you mean I have to walk?
Kamen wasn't thinking of Slashdot hackers. Good thing it can withstand 7 tons!
Just what America needs--one less way to exercise.
Let me elaborate a bit...
Taking a hash or SHA digest is bound to fail. One little character is off and the thing fails. That means customized emails, for example.
I'm working on a histogram based filter. You count the number of time a word occurrs in an email and you create a hash based upon the top twenty words and their occurance rate or so. If the spammer changes dear "bob" to dear "fred", bob and fred only occurr once or twice at most and are deemed insignificant by the algorithm and do not affect the hash. The more words you accept, the more accurate the fingerprint, although 20 or so seems to be accurate enough. Other configurable parts of the algorithm allow you to bump the word increment once if it occurs two times or three times instead of a one-to-one increment. Minute changes, therefore, will not affect the hash.
Furthermore, if an exact match is not made, you can keep compiling all histograms by sorting them into groups of like content, and then generate "master" general fingerprints, which can then be used with weights in a fuzzy algorithm to score a message for spammyness. Combined with a threshold (say 50% spamminess) you can decide whether or not you want to reject it. This again is only used if a direct match is not made. (If an exact fingerprint has not yet made it into the db).
This system still plays nice with distributed methods as you are still using a small hash code. If it were employed on one system, it would be easier to keep more detailed records of each hist; not just a hash.
Maybe one of these days I'll have a decent working prototype in python to share.
The best option would be a word count histogram filter. Then the spammer would have to entirely alter their language or sales pitch, which isn't going to happen. Just like handwriting, it is hard to change unless you make a whosale effort at changing it. They're too lazy, too.