Cheapasses is right. You get what you pay for, and if you're paying for impossible things, you should beware the catch. There is no such thing as unlimited bandwidth, or unlimited space...merely limits that you haven't hit yet.
A t1 line is still over $700 per month, so burstable bandwidth starts at more than $2 per gigabyte. People who are on better pipes pay way less, of course, but then again they need to maintain them, and technicians start at about $25 per hour. Servers need to be powered, backed up and maintained to prevent hackattacks. So when somebody offers you unlimited bandwidth, unlimited space, unlimited email with 24x7 support for a pretty number like $7.77 or $5.55 or whatever, they're basically lying to you.
Check your AUP. Somewhere in there you'll find a line saying that your unlimited bandwidth can be terminated at any time if you use too much of it. Unlimited really means "We're not telling you the limits. But you'll know when you hit them." Generally because your site takes off. You get popular, people start laughing at your jokes and caring about your weblog. Then your provider cuts the cord. Sucks, don't it?
See, ISPs at all levels make money by overselling. They tell you you have a T1, when really it's fractional. They tell you you have 256 kbit upstream, then it maxes at 192. The most egregious example of this is the El-Cheapo webhost, an animal I despised so much that I started my own crummy service to combat it. If you have the know-how, and you have the time, I suggest you do the same. It can be a lot of fun and offsets the cost of big web projects. Just don't harbor any dreams of getting stinko rich.
I remember the first time I had a site get "overnight popular." It was a certain web comic that we begged to come on board. In about two weeks ge went from moving 2 gig a month to over 50. And because we small timers get the short end of the bandwidth stick, his bill was about $200. Not his bill FROM us, but the bill TO us from our host for just his transfer. We didn't mark it up. That's a lot of money when you're a hobbiest. Shit, that was as much as we paid for everybody else's bandwidth that month.
We have a policy of not touching people's sites or restricting tranfer, but if we hadn't known the guy (and known he was good for the money, which his new fans donated in droves, we even threw in $30), we probably would have had to use the "no contract" clause and take the site offline. Damned if I'm paying for somebody else's popularity...
SOME money can be made by selling open products to savvy users. However, it's not that much. Certainly not as much as can be made by selling closed products in bulk to non-savvy users. We expect more for our $500, therefore we're a low margin market. Just ask VA Linux.
The more people who can get into that market, the lower prices will go. And so margin will decrease, and along with it the R&D budget which made this kickass formfactor possible in the first place.
End result is this: slightly increased convenience for Linux users who have some hard-on against binaries. Decreased prices. And no rewards whatsoever for the company that made it happen. No market advantage. You expect them to rely on the gratitude and karma from the teeny tiny savvy market? Via (+1, Interesting)?
What is it with the Linux community that so many people think that the hardware industry owes them a living...isn't it the other way 'round?
This is horseshit. Linux and BSD users are not a significant market share. By the very nature of the OS you aren't going to make much money off of sales...because even if you have a monopoly on that market segment, a lot of people just won't pay a fair price for your support. Ask any computer hobbiest about 3com (who have classically had excellent driver support) and the first words out of their mouth will probably be "too expensive." The reason it's too expensive is that it comes with awesome drivers.
Do they do more than you want them to do? Probably. So it's not worth it for you. But in arenas where 3com is the only player worthy to come to the plate (first release of BeOS comes to mind), they'd be fools to open their driver source. Because that will also open the market for clone hardware, and they will probably be forced to lower prices. They did the work. They made the product worthwhile. Why should they give all that up for your hobby?
The "competition could reverse engineer in a few days" comment is even more horseshit. If it takes your driver team three months to write something, then it won't take somebody else days to rip it apart. Not if they're doing things by clean room standards with hopes of making a legal competitor. Just writing the documentation so another team can write the driver will take as long as writing it in the first place. Which gives you a six month head start, if you're lucky. In the OSS world any information you can beg borrow or steal is perfectly all right, since you're never going to sell hardware based on that driver. So yeah, you can "feel in these lil holes." Competitors can't do that -- not unless they want to get sapped for their sales.
You're under the misapprehension that just because something CAN be done that it is USEFUL to do.
Releasing driver source code that can be compiled is like giving a person directions to your house. Releasing a driver they're forced to reverse engineer is giving them directions through 20 questions. Yeah, they might get there...but they'll have to work MUCH harder, and it will take MUCH logner, and there's a chance it will never work as well.
Why not make the competition work as hard as possible? It's better than invoking the DMCA, n'est-ce pas?
Most hardware companies use off the shelf parts. They aren't designing the technology so much as lciensing it and marketting it. The ONLY reason they are able to make money off of their products is that they have something that Generic Q. Solderinggun doesn't -- they have the ability to interface these hardware chips with a computer. And all of that magic happens in the driver.
I've seen lies here that drivers don't make money, and this is simply ludicrous. Let's take a real world example: back in 1997, both Iomega and Miro (later Pinnacle) marketted an MJPEG video input box based off of a Zoran chip. Zoran made a very very nice chip capable of massive resolutions, dozens of colour modes and bus mastering and all kinds of kick ass stuff. However, Iomega skimped on their drivers. The result was a product that was totally unable to operate at spec, because the driver had a fundamental flaw that prevented it from capturing at 29.976. Savvy video users quickly learned to cap at 30.10, drop a frame, and save $100+ over the cost of the similar Miro card. However, no matter how much the Slashdot community would like to think otherwise, you can't make money selling to ONLY savvy users. Iomega promptly dropped support, even though late model drivers were FIXING the issues. Miro, on the other hand, made money off of their superior drivers for years to come. Those drivers made that money.
If Miro opened the source of their drivers, GPL or otherwise, nothing would have stopped Iomega from getting them, modding them slightly to include their hardware, and releasing them back to the community. After all, they're not selling them. Good for everybody, right?
No good for Miro, whose dilligence in driver manufacture has just cost them countless sales. Their hardware is now just the sale as the other guy's but sells for much more.
Why the hell do you want hardware companies to lose money for your hobby? Are you so vain that you really think your 3% of the marketshare is worth that much to the VIAs of the world?
Why does a simple activation system make things harder? Well, define simple... does Windows XP have a simple activation system?
YES. Yes it is simple. My CEO, my coworkers, my mom, all were able to figure it out without documentation. Adding memory, removing memory, does not make it break. I have done both of these on my machine and it still runs. Quit your FUD, it's even less appealing from enemies of Microsoft than it is from big blue-screen themselves.
After all, it's not at all wierd for shareware vendors to sell you a license key on a website that depends on your name, your PC config, etc. So why is it naturally wierd for boxed software to require the same? After all, in this age of online software services, the box is often just advertisement anyway.
What you're describing -- purchasing a direct contract -- is EXACTLY what they're doing. The money exchanges hands through brick & mortar distributors, who consumers trust for some odd reason, and so you can pay cash, etc.
Yes...except that these 'solutions' don't solve the problem of people cracking a product and then selling it, without the support contract necessary for virus programs to stay up to date, at a massive product.
Because the end user STILL doesn't know that his software is not effective.
I don't see why a simple online activation system makes things "harder." It's certainly not harder than downloading and installing hacked software. So the only reason people who want protection from virus would actively seek out an illegal solution is if they were a) cheap or b) misinformed.
And if you're (a) cheap, there's always the free online scanner at trendmicro.com. I cleaned my mom's machine for no cost over VNC.
There are two things that bother me about people's misconceptions of Romeo and Juliet. The first is the line "Wherefore art thou, Romeo," and I can deal with it, because it's still better than people who don't understand the concept of contractions of homonyms.
The second is that people still think it's a heartwarming love story. It's not. It's a tragedy about the destruction of two feuding houses -- two families, too old to produce any more children, who lose all their kin as a result of this pairing. Whether lovesick Romeo (who immediately forgets the girl he had previously pined for when he meets a new chick who is WAY too young for him) and dutiful Juliet are truly in love is up to the interpretter, though I'd have to say in true deconstructionist spirit that there's no way. But if it were really about their "love," the story would have ended with their deaths. It continues...not merely as denoument, but to deliver the message: foolish feuding only leads to tragedy.
It's an anti-mob play, same as Julius Caeser. There's nothing romantic in the second act at all, just a lot of foolish teenage sentiment and real adult grief. Therefore, a phrase like "we're starcrossed lovers, like Romeo and Juliet," has always seemed insultingly ironic to me. Except when Milhouse said it.
Dictation programs have replaced secreataries and typists.
Tools like Google, SQL and mapping software do a better job of researching information than people do.
Machines perform very well in tasks where we boss them around. They don't perform equally well when they have to perform a lot of decision making. This is an attempt to bring them to a more passable level. And since technology is always replacing people, I think designing technology with the vision of augmenting a person's computer usage is very noble. And it's something that's very important to point out when we've got doom and gloom pundits everywhere.
Acutally, I'd prefer a spam filter that was 90% successful, if there was no chance that it would also destroy something I didn't want. You know, like when Snopes blacklists somethingawful.com. Very effective at eliminating things, but not very deterministic.
Anyway, don't think of this as an automated process like a spam filter. This is more like an electronic co-worker, in that it learns how you would do things and then you can tell it to do them again for you. Except I've never had one that was within 90% of my stunted vision. Having one of these around to rerun tests for me and set up config files would save me a pantload of time.
Um, the elevators in H2G2 weren't polite. They were clinically depressed, on account of they were precognitive and could see the future and it was pretty bleak.
It was the computer that was polite. Doors were self satisfied (it is their pleasure to open for you, and their pleasure to close again with the satisfaction of a job well done).
Whoa. Air Traffic Controllers know where their shit is. They know how to use their machines.
And if they have to be clicking through directories to tell where to save the document they were working on when they have to pull up an emergency response program, that's taking time.
A software agent that learns "when there's an emergency, save X type of document in directory Y and pull up program Z" saves the grunt work in something the controller already knows how to do. You and I probably do this now through macros, scripts, etc. You can't possibly expect every user on the planet to learn how to perform complicated conditional logic in a programming language...it's a waste of time and resources.
Something like this...that learns macros on its own, and gives the user the option to perform them...would be a godsend, even for me. Because sometimes I get lazy...like when I'm running five scripts to convert a SQL statement from one language to another and don't want to take an hour to write a script to run them in succession based on conditions in the converted base because the statements shouldn't even NEED to be converted except SOME people don't know how to use a fucking database...a program that, after I did something complex twice, gave me the option to do it a third time (and pop up monster.com as well), would be awesome.
I think it's supposed to mean "fewer instructions to perform the same task." It could also mean stronger bitstrength. Your 1-1000 example was kind of crap...try it again with 1,000,000,000,000 and you get 60 states for base 2 vs 39 for base 3, a pretty good savings. In 64 bit computing, your max int in base 2 is 18446744073709551616. Your max int in base 3 is 3433683820292512484657849089281. A 32 bit ternary system would have 1853020188851841 distinct values, which is only 10,000 off from the 64 bit binary! Meaning greater precision with less storage for roughly equivalent precision, all in a system with greater reliablity and less complexity.
That's a good idea. Now, the trouble comes when you realize that ternary logic is much more complex than binary, resulting in more circuitry per bit. It requires more and cleaner power with more switches. I have to wager that in the end it's not worth the hassle. But it would put us in good with the aliens from Rama, the Robotech Masters, etc.
Except that you'd need twice as many lines and switches to do this shit. Because -3.3V moves BACKWARDS.
Re:I like the indeterminant ternary logic concept
on
Beyond Binary Computing?
·
· Score: 3, Informative
The problem with this is that fuzzy logic is no more ternary than it is in binary. Fuzzy logic is about effectively weighing options. With ternary logic, you've only increased your options by 50%. Fuzzy logic is generally probabalistic...which means it's nicest when utilized with sufficiently large integers or, more importantly, huge floating point numbers. Like a nice 64 bit quad.
Consider a typical "fuzzy" logic problem of when to stop a car. You want to weigh variables like the speed of a car, the amount of force to apply, distance to the stopping point and other options like the existance of pedestrians (generally, if there's people within 20 feet of your stopping point, like say at a bus stop near a stop sign, you'll want to slow down quicker at first but roll longer, to decrease the effect of a fluke accident).
Lots of variables. Lots of choice. Lots of probability to weigh. Having an extra option out of 3 does not help you. Having 64 bits to work with does.
I have a hard time coming up with problems in my line of software dev in which ternary or quaternary logic is any more useful than nested binary logic or some fun probability and calculus. Mostly because it's rare that I care about anything other than STOP or GO, ON or OFF. About the only time I do care is when I'm dealing with a database (YES, NO or NULL [no data])...all the rest of the time, alternative options are best handled with an enumerated type or a nice exception.
Anyway, it's all well and good to talk about ternary computing being 'faster' with less overhead, but it's never really going to take off. It will take at least an extra year to train engineers to use the new logic effectively and for them to learn the tricks...and in that year, binary computing will have doubled. And when you live in a world where most software isn't optimized anyway...waiting for a slightly faster logic system that 9/10 of programmers will merely treat as binary because it's easier to understand in more comfortable is a waste of everyone's resources.
Even the terminology is bad. True, false, alpha. Ugh. If P is sort of true, then kind of do q.
To be honest, I prefer the latest incarnation of Rick Jones as the magically linked alter ego of omnipowerful insane superhero Captain Marvel to some wussy "Lifetime network" Rick Jones.
Oh what, you're not reading Peter David's run on Captain Marvel? You should be. It's the most amazing what-if morality play ever published by Marvel. Pick up last month's jaded, thinly veiled allegory of modern Bush imperialism. Or read issues 1 through 7, in which absolute power drives him insane, so he learns heroics from the punisher, spends time as the champion of the kree and finally winds up destroying the universe (just cos it's something to do). And in the meantime, Rick Jones has become a rock star in the Microverse.
Here, here. I'm of the opinion that it's the Fictional Science which creates the enjoyment. It's all a matter of what-if.
Hulk is a perfect example. When he was created, everybody was scared of nuclear weapons, because they were powerful and mysterious. Marvel said, "What if a gamma bomb were able to create monsters and in doing so updates Jeckyll and Hyde?"
Good science? No. Of course not. Good science would have our man Banner dead from radiation sickness and buried in a lead lined coffin. The story is rather short and tragic. Now, with this fictional, impossible, fantastic science, Hulk is an interesting character and a symbol of inner conflict.
The "Science" in science fiction is crap because it makes the stories more interesting. Complaining that it's crap is missing the point entirely -- it's like complaining that conceits are unrealistic, something that Willy Shakespearre already touched on..."My mistress' eyes are nothing like the sun," etc.
I do think there's a trade off, and the best science fiction adheres as closely to "real world" physics, chemistry and biology as it can. But you have to excuse where it steps off, or accept some VERY boring shit:
"Warp Factor 5, Mr. Sulu."
"Ahh, but captain, warp dynamics violate general relativity, and therefore are bad science. Besides which, it does not make sense that they are measured in factors when those factors have decimal values."
"I guess we'll just float around here for a while, then. Maybe I'll make out with a blonde crew member, being as there's apparently no sexual harrassment in space."
Virtual users do not really need to have a local account at all...they need to have an entry in the users database, and that by default uses local accounts. But if you want to make a big directory full of mail drop folders and make everything perfectly virtual...then qmail allows you to do that.
You will need a UNIX user account for your ploy. But really, you need a UNIX user account to do anything. I made one called virtualusers, which has no permissions except in/home/mailtricks. Then I did this in/var/qmail/users/assign: =nosuchuser:virtualusers :::/home/mailtricks/nosuch user::: =anotheruser:virtualusers:::/home/mailtri cks/anoth eruser:::
virtualusers is used as the accessing UNIX user for mail delivery. As long as it has rwx access to/home/mailtricks, any number of "users" can use the same UNIX account. The cool thing is, if you use qmail-pop3 and a qmail compatible imap server (like courier), they can use the same assign database for checking mail as you use to deliver it.
All you have to do is manage the virtualdomains and assign files, which isn't so hard. I'm working on a jsp web admin system for virtual users which does exactly that...I'll be releasing it under BSD if it ever matures.
Oh shut up. I capitalized qmail because it was at the beginning of the sentance. Or are you claiming that Dan Bernstein is actually e e cummings and I should drop all grammatical rules when referencing him?
Slashdot Mirror
Yes, lucky for him there are 2200 shift, alt, control and apple keys!
Well, no. It'd be deep BONDI blue, if it were 1997 (hint: it's not). Nowadays, it's Deep Titanium or Deep Snow, something Virginia has never seen.
WTF are you talking about? Everything we do is over SSH. SSL, on the otherhand, is expensive or untrusted. We have the untrusted version available.
Cheapasses is right. You get what you pay for, and if you're paying for impossible things, you should beware the catch. There is no such thing as unlimited bandwidth, or unlimited space...merely limits that you haven't hit yet.
A t1 line is still over $700 per month, so burstable bandwidth starts at more than $2 per gigabyte. People who are on better pipes pay way less, of course, but then again they need to maintain them, and technicians start at about $25 per hour. Servers need to be powered, backed up and maintained to prevent hackattacks. So when somebody offers you unlimited bandwidth, unlimited space, unlimited email with 24x7 support for a pretty number like $7.77 or $5.55 or whatever, they're basically lying to you.
Check your AUP. Somewhere in there you'll find a line saying that your unlimited bandwidth can be terminated at any time if you use too much of it. Unlimited really means "We're not telling you the limits. But you'll know when you hit them." Generally because your site takes off. You get popular, people start laughing at your jokes and caring about your weblog. Then your provider cuts the cord. Sucks, don't it?
See, ISPs at all levels make money by overselling. They tell you you have a T1, when really it's fractional. They tell you you have 256 kbit upstream, then it maxes at 192. The most egregious example of this is the El-Cheapo webhost, an animal I despised so much that I started my own crummy service to combat it. If you have the know-how, and you have the time, I suggest you do the same. It can be a lot of fun and offsets the cost of big web projects. Just don't harbor any dreams of getting stinko rich.
I remember the first time I had a site get "overnight popular." It was a certain web comic that we begged to come on board. In about two weeks ge went from moving 2 gig a month to over 50. And because we small timers get the short end of the bandwidth stick, his bill was about $200. Not his bill FROM us, but the bill TO us from our host for just his transfer. We didn't mark it up. That's a lot of money when you're a hobbiest. Shit, that was as much as we paid for everybody else's bandwidth that month.
We have a policy of not touching people's sites or restricting tranfer, but if we hadn't known the guy (and known he was good for the money, which his new fans donated in droves, we even threw in $30), we probably would have had to use the "no contract" clause and take the site offline. Damned if I'm paying for somebody else's popularity...
SOME money can be made by selling open products to savvy users. However, it's not that much. Certainly not as much as can be made by selling closed products in bulk to non-savvy users. We expect more for our $500, therefore we're a low margin market. Just ask VA Linux.
The more people who can get into that market, the lower prices will go. And so margin will decrease, and along with it the R&D budget which made this kickass formfactor possible in the first place.
End result is this: slightly increased convenience for Linux users who have some hard-on against binaries. Decreased prices. And no rewards whatsoever for the company that made it happen. No market advantage. You expect them to rely on the gratitude and karma from the teeny tiny savvy market? Via (+1, Interesting)?
What is it with the Linux community that so many people think that the hardware industry owes them a living...isn't it the other way 'round?
This is horseshit. Linux and BSD users are not a significant market share. By the very nature of the OS you aren't going to make much money off of sales...because even if you have a monopoly on that market segment, a lot of people just won't pay a fair price for your support. Ask any computer hobbiest about 3com (who have classically had excellent driver support) and the first words out of their mouth will probably be "too expensive." The reason it's too expensive is that it comes with awesome drivers.
Do they do more than you want them to do? Probably. So it's not worth it for you. But in arenas where 3com is the only player worthy to come to the plate (first release of BeOS comes to mind), they'd be fools to open their driver source. Because that will also open the market for clone hardware, and they will probably be forced to lower prices. They did the work. They made the product worthwhile. Why should they give all that up for your hobby?
The "competition could reverse engineer in a few days" comment is even more horseshit. If it takes your driver team three months to write something, then it won't take somebody else days to rip it apart. Not if they're doing things by clean room standards with hopes of making a legal competitor. Just writing the documentation so another team can write the driver will take as long as writing it in the first place. Which gives you a six month head start, if you're lucky. In the OSS world any information you can beg borrow or steal is perfectly all right, since you're never going to sell hardware based on that driver. So yeah, you can "feel in these lil holes." Competitors can't do that -- not unless they want to get sapped for their sales.
That they are not playing them loud enough. I played Silent Hill 3 a week ago and I still can't hear any sound over 10,000 Hz.
Don't worry about me, though. I broke out my Mark Lanegan CDs. It's gonna be okay.
You're under the misapprehension that just because something CAN be done that it is USEFUL to do.
Releasing driver source code that can be compiled is like giving a person directions to your house. Releasing a driver they're forced to reverse engineer is giving them directions through 20 questions. Yeah, they might get there...but they'll have to work MUCH harder, and it will take MUCH logner, and there's a chance it will never work as well.
Why not make the competition work as hard as possible? It's better than invoking the DMCA, n'est-ce pas?
How are they doing themselves a favor?
Most hardware companies use off the shelf parts. They aren't designing the technology so much as lciensing it and marketting it. The ONLY reason they are able to make money off of their products is that they have something that Generic Q. Solderinggun doesn't -- they have the ability to interface these hardware chips with a computer. And all of that magic happens in the driver.
I've seen lies here that drivers don't make money, and this is simply ludicrous. Let's take a real world example: back in 1997, both Iomega and Miro (later Pinnacle) marketted an MJPEG video input box based off of a Zoran chip. Zoran made a very very nice chip capable of massive resolutions, dozens of colour modes and bus mastering and all kinds of kick ass stuff. However, Iomega skimped on their drivers. The result was a product that was totally unable to operate at spec, because the driver had a fundamental flaw that prevented it from capturing at 29.976. Savvy video users quickly learned to cap at 30.10, drop a frame, and save $100+ over the cost of the similar Miro card. However, no matter how much the Slashdot community would like to think otherwise, you can't make money selling to ONLY savvy users. Iomega promptly dropped support, even though late model drivers were FIXING the issues. Miro, on the other hand, made money off of their superior drivers for years to come. Those drivers made that money.
If Miro opened the source of their drivers, GPL or otherwise, nothing would have stopped Iomega from getting them, modding them slightly to include their hardware, and releasing them back to the community. After all, they're not selling them. Good for everybody, right?
No good for Miro, whose dilligence in driver manufacture has just cost them countless sales. Their hardware is now just the sale as the other guy's but sells for much more.
Why the hell do you want hardware companies to lose money for your hobby? Are you so vain that you really think your 3% of the marketshare is worth that much to the VIAs of the world?
Why does a simple activation system make things harder? Well, define simple... does Windows XP have a simple activation system?
YES. Yes it is simple. My CEO, my coworkers, my mom, all were able to figure it out without documentation. Adding memory, removing memory, does not make it break. I have done both of these on my machine and it still runs. Quit your FUD, it's even less appealing from enemies of Microsoft than it is from big blue-screen themselves.
After all, it's not at all wierd for shareware vendors to sell you a license key on a website that depends on your name, your PC config, etc. So why is it naturally wierd for boxed software to require the same? After all, in this age of online software services, the box is often just advertisement anyway.
What you're describing -- purchasing a direct contract -- is EXACTLY what they're doing. The money exchanges hands through brick & mortar distributors, who consumers trust for some odd reason, and so you can pay cash, etc.
Yes...except that these 'solutions' don't solve the problem of people cracking a product and then selling it, without the support contract necessary for virus programs to stay up to date, at a massive product.
Because the end user STILL doesn't know that his software is not effective.
I don't see why a simple online activation system makes things "harder." It's certainly not harder than downloading and installing hacked software. So the only reason people who want protection from virus would actively seek out an illegal solution is if they were a) cheap or b) misinformed.
And if you're (a) cheap, there's always the free online scanner at trendmicro.com. I cleaned my mom's machine for no cost over VNC.
There are two things that bother me about people's misconceptions of Romeo and Juliet. The first is the line "Wherefore art thou, Romeo," and I can deal with it, because it's still better than people who don't understand the concept of contractions of homonyms.
The second is that people still think it's a heartwarming love story. It's not. It's a tragedy about the destruction of two feuding houses -- two families, too old to produce any more children, who lose all their kin as a result of this pairing. Whether lovesick Romeo (who immediately forgets the girl he had previously pined for when he meets a new chick who is WAY too young for him) and dutiful Juliet are truly in love is up to the interpretter, though I'd have to say in true deconstructionist spirit that there's no way. But if it were really about their "love," the story would have ended with their deaths. It continues...not merely as denoument, but to deliver the message: foolish feuding only leads to tragedy.
It's an anti-mob play, same as Julius Caeser. There's nothing romantic in the second act at all, just a lot of foolish teenage sentiment and real adult grief. Therefore, a phrase like "we're starcrossed lovers, like Romeo and Juliet," has always seemed insultingly ironic to me. Except when Milhouse said it.
Giving advice constantly will inevitably lead to wrong and/or unwanted advice at one point.
You're married, aren't you?
Why is it laughable?
Robots have replaced workers in factories.
Dictation programs have replaced secreataries and typists.
Tools like Google, SQL and mapping software do a better job of researching information than people do.
Machines perform very well in tasks where we boss them around. They don't perform equally well when they have to perform a lot of decision making. This is an attempt to bring them to a more passable level. And since technology is always replacing people, I think designing technology with the vision of augmenting a person's computer usage is very noble. And it's something that's very important to point out when we've got doom and gloom pundits everywhere.
Acutally, I'd prefer a spam filter that was 90% successful, if there was no chance that it would also destroy something I didn't want. You know, like when Snopes blacklists somethingawful.com. Very effective at eliminating things, but not very deterministic.
Anyway, don't think of this as an automated process like a spam filter. This is more like an electronic co-worker, in that it learns how you would do things and then you can tell it to do them again for you. Except I've never had one that was within 90% of my stunted vision. Having one of these around to rerun tests for me and set up config files would save me a pantload of time.
Um, the elevators in H2G2 weren't polite. They were clinically depressed, on account of they were precognitive and could see the future and it was pretty bleak.
It was the computer that was polite. Doors were self satisfied (it is their pleasure to open for you, and their pleasure to close again with the satisfaction of a job well done).
Whoa. Air Traffic Controllers know where their shit is. They know how to use their machines.
And if they have to be clicking through directories to tell where to save the document they were working on when they have to pull up an emergency response program, that's taking time.
A software agent that learns "when there's an emergency, save X type of document in directory Y and pull up program Z" saves the grunt work in something the controller already knows how to do. You and I probably do this now through macros, scripts, etc. You can't possibly expect every user on the planet to learn how to perform complicated conditional logic in a programming language...it's a waste of time and resources.
Something like this...that learns macros on its own, and gives the user the option to perform them...would be a godsend, even for me. Because sometimes I get lazy...like when I'm running five scripts to convert a SQL statement from one language to another and don't want to take an hour to write a script to run them in succession based on conditions in the converted base because the statements shouldn't even NEED to be converted except SOME people don't know how to use a fucking database...a program that, after I did something complex twice, gave me the option to do it a third time (and pop up monster.com as well), would be awesome.
Amateur. You have two joints on all your fingers and three distinct states -- up, bent, and down.
That's ternary. And it means that instead of 1023, you can count up to 59049.
Which is great for counting pennies before buying an iPod.
I think it's supposed to mean "fewer instructions to perform the same task." It could also mean stronger bitstrength. Your 1-1000 example was kind of crap...try it again with 1,000,000,000,000 and you get 60 states for base 2 vs 39 for base 3, a pretty good savings. In 64 bit computing, your max int in base 2 is 18446744073709551616. Your max int in base 3 is 3433683820292512484657849089281. A 32 bit ternary system would have 1853020188851841 distinct values, which is only 10,000 off from the 64 bit binary! Meaning greater precision with less storage for roughly equivalent precision, all in a system with greater reliablity and less complexity.
That's a good idea. Now, the trouble comes when you realize that ternary logic is much more complex than binary, resulting in more circuitry per bit. It requires more and cleaner power with more switches. I have to wager that in the end it's not worth the hassle. But it would put us in good with the aliens from Rama, the Robotech Masters, etc.
Except that you'd need twice as many lines and switches to do this shit. Because -3.3V moves BACKWARDS.
The problem with this is that fuzzy logic is no more ternary than it is in binary. Fuzzy logic is about effectively weighing options. With ternary logic, you've only increased your options by 50%. Fuzzy logic is generally probabalistic...which means it's nicest when utilized with sufficiently large integers or, more importantly, huge floating point numbers. Like a nice 64 bit quad.
Consider a typical "fuzzy" logic problem of when to stop a car. You want to weigh variables like the speed of a car, the amount of force to apply, distance to the stopping point and other options like the existance of pedestrians (generally, if there's people within 20 feet of your stopping point, like say at a bus stop near a stop sign, you'll want to slow down quicker at first but roll longer, to decrease the effect of a fluke accident).
Lots of variables. Lots of choice. Lots of probability to weigh. Having an extra option out of 3 does not help you. Having 64 bits to work with does.
I have a hard time coming up with problems in my line of software dev in which ternary or quaternary logic is any more useful than nested binary logic or some fun probability and calculus. Mostly because it's rare that I care about anything other than STOP or GO, ON or OFF. About the only time I do care is when I'm dealing with a database (YES, NO or NULL [no data])...all the rest of the time, alternative options are best handled with an enumerated type or a nice exception.
Anyway, it's all well and good to talk about ternary computing being 'faster' with less overhead, but it's never really going to take off. It will take at least an extra year to train engineers to use the new logic effectively and for them to learn the tricks...and in that year, binary computing will have doubled. And when you live in a world where most software isn't optimized anyway...waiting for a slightly faster logic system that 9/10 of programmers will merely treat as binary because it's easier to understand in more comfortable is a waste of everyone's resources.
Even the terminology is bad. True, false, alpha. Ugh. If P is sort of true, then kind of do q.
To be honest, I prefer the latest incarnation of Rick Jones as the magically linked alter ego of omnipowerful insane superhero Captain Marvel to some wussy "Lifetime network" Rick Jones.
Oh what, you're not reading Peter David's run on Captain Marvel? You should be. It's the most amazing what-if morality play ever published by Marvel. Pick up last month's jaded, thinly veiled allegory of modern Bush imperialism. Or read issues 1 through 7, in which absolute power drives him insane, so he learns heroics from the punisher, spends time as the champion of the kree and finally winds up destroying the universe (just cos it's something to do). And in the meantime, Rick Jones has become a rock star in the Microverse.
Good comic, that.
Here, here. I'm of the opinion that it's the Fictional Science which creates the enjoyment. It's all a matter of what-if.
Hulk is a perfect example. When he was created, everybody was scared of nuclear weapons, because they were powerful and mysterious. Marvel said, "What if a gamma bomb were able to create monsters and in doing so updates Jeckyll and Hyde?"
Good science? No. Of course not. Good science would have our man Banner dead from radiation sickness and buried in a lead lined coffin. The story is rather short and tragic. Now, with this fictional, impossible, fantastic science, Hulk is an interesting character and a symbol of inner conflict.
The "Science" in science fiction is crap because it makes the stories more interesting. Complaining that it's crap is missing the point entirely -- it's like complaining that conceits are unrealistic, something that Willy Shakespearre already touched on..."My mistress' eyes are nothing like the sun," etc.
I do think there's a trade off, and the best science fiction adheres as closely to "real world" physics, chemistry and biology as it can. But you have to excuse where it steps off, or accept some VERY boring shit:
"Warp Factor 5, Mr. Sulu."
"Ahh, but captain, warp dynamics violate general relativity, and therefore are bad science. Besides which, it does not make sense that they are measured in factors when those factors have decimal values."
"I guess we'll just float around here for a while, then. Maybe I'll make out with a blonde crew member, being as there's apparently no sexual harrassment in space."
Virtual users do not really need to have a local account at all...they need to have an entry in the users database, and that by default uses local accounts. But if you want to make a big directory full of mail drop folders and make everything perfectly virtual...then qmail allows you to do that.
/home/mailtricks. Then I did this in /var/qmail/users/assign:s :::/home/mailtricks/nosuch user:::i cks/anoth eruser:::
/home/mailtricks, any number of "users" can use the same UNIX account. The cool thing is, if you use qmail-pop3 and a qmail compatible imap server (like courier), they can use the same assign database for checking mail as you use to deliver it.
You will need a UNIX user account for your ploy. But really, you need a UNIX user account to do anything. I made one called virtualusers, which has no permissions except in
=nosuchuser:virtualuser
=anotheruser:virtualusers:::/home/mailtr
virtualusers is used as the accessing UNIX user for mail delivery. As long as it has rwx access to
All you have to do is manage the virtualdomains and assign files, which isn't so hard. I'm working on a jsp web admin system for virtual users which does exactly that...I'll be releasing it under BSD if it ever matures.
Oh shut up. I capitalized qmail because it was at the beginning of the sentance. Or are you claiming that Dan Bernstein is actually e e cummings and I should drop all grammatical rules when referencing him?