Well, you can, but it would have to be tailored to the data, meaning the information is being kept in the algorithm... and at that point it's an algorithm plus its own store of data... and hasn't reduced the required information by much. This is a completely inescapable factor of lossless compression.
You know, that's actually a poor ratio given the probable fact that there are more duplications of chemicals than there are actual different chemicals involved. Then again, I didn't exactly frob the switches controlling the ratio.
I wonder if there are any algorithms for determining information density... there's gotta be.
I agree with the spirit of what you say, but I'd really like you to back up the bit about mailsystems keeping copies. I'm sure many do, but certainly not most.
I happen to do maintenance on over 200 mailservers (different companies) and not one of them retains messages beyond normal queuing. In fact, WE DON'T WANT YOUR MAIL ON OUR SERVERS. Think about the average email volume per day.
Now, mutiply that by the number of days you'd expect this mail to be retained... that's a lot of wasted money on storage for something we're not even liable for.
I've written a few in Perl for quick breakdowns. No graphing, but the equivalent textual information. I recall seeing a few projects of a similar nature on sweetcode, with graphing (3D?) unfortunately I don't recall what they might have been called.
This could be enforced by the ISP. In fact, it could be standardized throughout, built into SMTPvX or whatever. IF there is an RFC on it, for example, and people start making servers for it, clients will be written, eventally SMTP as we know it could be phased out entirely. Of course, this means that it couldn't be fixed over night.
A two second sleep would kill an ISP. It would do a lot of harm to a spammer, but it would have virtually no effect on legitimate clients. As such, it's best to enforce it on the client end so that it harms only people sending large volumes of mail. Spammers get virtually nothing per message which means they have to have massive lists to make it worth the effort.
If spammers can compute fast enough to make a given level ineffective, then up the requirement by 4 or 5 digits, making the processing take 16-32 times as long.
Um. A 2 second sleep would put you out of business, and have zero effect on spam. Spammers tend to use their own servers, or zombies to send spam. I catch spammers spammers regularly, and they're always running their own mailer software, which establishes its own connections with recipient servers. That's not exactly the kind of thing you can throttle.
If you're an ISP and 98% of the incoming mail is delayed for 4 hours, and 98% of that is being delayed for another 4 hours and so on, then most of your mail will get dropped after 5 days. I'm not going to bother showing the math; the point is you'll lose most of your customers and won't be in business long enough because the only email people will be getting, for the most part, will be their spam.
As I see it, what's to stop you from taking a POW, cashing it in for an RPOW, then repeating the process over and over for a bit... over the length of your entire spam-to list, then using them?
Sure, they're in the "seen" database, but the point of the (R)POW is to be intrinsically verifiable, without having to consult a database.
Unless the server is going to take its time in getting you a replacement RPOW, I don't see how this is actually useful vs. spam. I mean, if it's relatively quick to issue a new RPOW, then wouldn't the time involved in accumulating the spam list be greater overall?
Finally, how many bits should an authorized email cost?
This sounds incredible, but it's true. We had a power supply that was so broken that it made a noise while unplugged and turned off. This was a bit frightening, but not only did it not work, but plugging it in and turning it on would cause the noise to stop.
Having a patently broken device, we decided to throw it against the cement floow, and apply hammers to the damn thing. When we were finally told to stop making such a noise, we ended the madness. I think Eris was in the building at the time, because not only had it stopped making noise while unplugged, but when we connected it to the machine and plugged it in, it worked!
Try being nice to your potential client base *before* you ask them for money. Like, shut off the ads first, then offer features that a) count, and b) can't be gotten with a no-brainer filtering proxy server.
You're right. We should stop experimenting with code altogether.
Screw that.
I've talked with these people on mailing lists and otherwise... they know a lot about computation, and for good reason. They have a powerful command of code in general, I find, and can answer just about any conceptual programming related questions. They're smart, and their work on these languages only hones their abilities... it's like they've created their own education system for themselves, just because you can't appreciate or understand it doesn't mean it's pointless.
This has been discussed from time to time on the Cat's Eye Tech. mailing list (now defunct) and on lang@ and misc@esoteric.sange.fi (successor to the dead list.)... but nothing was ever done about it. There was always more interest in dreaming up new paradigms than actually writing software with them.
Well, an exception is the OS written in Brainfuck... but I'll leave it as an excercise for you guys to find it...
Heh, I've been hanging out on mailing lists with the creator of said language. He's a really neat guy, and the other denizens of the list are sharp, too.
If you're really interested, you might look at www.catseye.mb.ca, which is Chris Pressey's (creator of befunge) website.
Slashdot Mirror
Different combinations depending on various factors.
Remember, ^G makes noise on the terminal.
Just create a named pipe connected to an app that spits out a random wave file on access.
Oh, right, you can't do that and use Outlook at the same time.
is Slashdot.
Yeah, but it's pretty obvious he didn't read the results of the search. There is stuff there. Much to my dismay.
I can't believe that Ask Slashdot caters to people who are too lazy to search Freshmeat.
Don't I wish.
Well, you can, but it would have to be tailored to the data, meaning the information is being kept in the algorithm... and at that point it's an algorithm plus its own store of data... and hasn't reduced the required information by much. This is a completely inescapable factor of lossless compression.
You know, that's actually a poor ratio given the probable fact that there are more duplications of chemicals than there are actual different chemicals involved. Then again, I didn't exactly frob the switches controlling the ratio.
I wonder if there are any algorithms for determining information density... there's gotta be.
But it is highly redundant:
$ du -sh chem
68k chem
$ gzip chem
$ du -sh chem.gz
12k chem.gz
Don't you know that the hackish word for GOD is 'RNG'?! You'd best _ for forgiveness.
Okay, that makes more sense.
I agree with the spirit of what you say, but I'd really like you to back up the bit about mailsystems keeping copies. I'm sure many do, but certainly not most.
I happen to do maintenance on over 200 mailservers (different companies) and not one of them retains messages beyond normal queuing. In fact, WE DON'T WANT YOUR MAIL ON OUR SERVERS. Think about the average email volume per day.
Now, mutiply that by the number of days you'd expect this mail to be retained... that's a lot of wasted money on storage for something we're not even liable for.
I've written a few in Perl for quick breakdowns. No graphing, but the equivalent textual information. I recall seeing a few projects of a similar nature on sweetcode, with graphing (3D?) unfortunately I don't recall what they might have been called.
Haha, good show.
> Can you pass this test?
Yes. But:
> C#.
You can't pay me enough.
This could be enforced by the ISP. In fact, it could be standardized throughout, built into SMTPvX or whatever. IF there is an RFC on it, for example, and people start making servers for it, clients will be written, eventally SMTP as we know it could be phased out entirely. Of course, this means that it couldn't be fixed over night.
A two second sleep would kill an ISP. It would do a lot of harm to a spammer, but it would have virtually no effect on legitimate clients. As such, it's best to enforce it on the client end so that it harms only people sending large volumes of mail. Spammers get virtually nothing per message which means they have to have massive lists to make it worth the effort.
If spammers can compute fast enough to make a given level ineffective, then up the requirement by 4 or 5 digits, making the processing take 16-32 times as long.
Um. A 2 second sleep would put you out of business, and have zero effect on spam. Spammers tend to use their own servers, or zombies to send spam. I catch spammers spammers regularly, and they're always running their own mailer software, which establishes its own connections with recipient servers. That's not exactly the kind of thing you can throttle.
If you're an ISP and 98% of the incoming mail is delayed for 4 hours, and 98% of that is being delayed for another 4 hours and so on, then most of your mail will get dropped after 5 days. I'm not going to bother showing the math; the point is you'll lose most of your customers and won't be in business long enough because the only email people will be getting, for the most part, will be their spam.
As I see it, what's to stop you from taking a POW, cashing it in for an RPOW, then repeating the process over and over for a bit... over the length of your entire spam-to list, then using them?
Sure, they're in the "seen" database, but the point of the (R)POW is to be intrinsically verifiable, without having to consult a database.
Unless the server is going to take its time in getting you a replacement RPOW, I don't see how this is actually useful vs. spam. I mean, if it's relatively quick to issue a new RPOW, then wouldn't the time involved in accumulating the spam list be greater overall?
Finally, how many bits should an authorized email cost?
This sounds incredible, but it's true. We had a power supply that was so broken that it made a noise while unplugged and turned off. This was a bit frightening, but not only did it not work, but plugging it in and turning it on would cause the noise to stop.
Having a patently broken device, we decided to throw it against the cement floow, and apply hammers to the damn thing. When we were finally told to stop making such a noise, we ended the madness. I think Eris was in the building at the time, because not only had it stopped making noise while unplugged, but when we connected it to the machine and plugged it in, it worked!
Hypertaurus?
Desperate for subscribers?
Try being nice to your potential client base *before* you ask them for money. Like, shut off the ads first, then offer features that a) count, and b) can't be gotten with a no-brainer filtering proxy server.
It worked for K5 you know.
All code under the GPL is copyrighted (owned) by the person (or group, or organisation) that wrote the code.
Actually, you're urged to sign over the rights to the FSF. It's not a requirement, but it's supposedly helps the GPL stand up in court.
Yeah, of course that would be horribly inefficient. I mean, you could do this instead:
"H","e","l","l","o",","," ","w","o","r","l","d",".",55+,@
but that would be nasty. I'd sooner approach from the east, and read it backwards, but read right-to-left instead... that's a lot simpler.
You're right. We should stop experimenting with code altogether.
Screw that.
I've talked with these people on mailing lists and otherwise... they know a lot about computation, and for good reason. They have a powerful command of code in general, I find, and can answer just about any conceptual programming related questions. They're smart, and their work on these languages only hones their abilities... it's like they've created their own education system for themselves, just because you can't appreciate or understand it doesn't mean it's pointless.
This has been discussed from time to time on the Cat's Eye Tech. mailing list (now defunct) and on lang@ and misc@esoteric.sange.fi (successor to the dead list.)... but nothing was ever done about it. There was always more interest in dreaming up new paradigms than actually writing software with them.
Well, an exception is the OS written in Brainfuck... but I'll leave it as an excercise for you guys to find it...
Heh, I've been hanging out on mailing lists with the creator of said language. He's a really neat guy, and the other denizens of the list are sharp, too.
If you're really interested, you might look at www.catseye.mb.ca, which is Chris Pressey's (creator of befunge) website.