And increase in cpu required is somewhat offset by reduced storage space needed for spam.
This is true, in that better spam filtering will allow the ISP to put off buying bigger disks for a while. But a crucial difference between disk space and CPU cycles is that you can reuse disk, but you can't reuse CPU.
From the ISP's point of view, it may not make much sense to spend, say, 5 microseconds to run a Bayesian filter on a 5K message, and then decide to dump it in/dev/null instead of/var/spool/mail, when half an hour later, the mailbox owner will read the subject line and delete the message anyway. Either way, the ISP has the 5K back, but the 5 microseconds are gone forever.
Of course, the mailbox owner may well take the view that he'd rather have the computer waste 5 microseconds deleting a message than him waste 5 whole seconds doing it himself. This is what angers me most about spam, more than the fact that it's driving up my costs: these bastards are stealing my time, and giving me nothing in return.
If spam is really a big problem, then facilitating individual customers' filtering is a selling point.
Absolutely. I'm getting near the point where I'd pay for something like that. But at the back of my mind is a little voice asking: "Why should you have to pay not to receive crap that has no benefit for you and that you never asked for in the first place?"
If the ISPs won't help, we simply need to get client-side filtering into the hands of as many people, usable, and used as much as possible, reducing the spammers' audience.
Anything that improves the signal-to-noise ratio in peoples' inboxes is good, but I don't believe client-side filtering will make much of a dent in the spammers' market. The whole point of spam is that a spammer can afford to send a million messages and make maybe ten sales. Anybody who hates spam enough to install a client-side filter is probably already one of the 999,990 people who wouldn't have fallen for the scam even if they had read the message. But if the result is that a person's inbox becomes useable again, would they care how it happened?
Better yet, if this law really is as broad as claimed, most software for sending spam is now illegal in Michigan. It might even be illegal to operate an open relay, or to use one for spamming. Persuading law enforcement to go after a spammer on these grounds would be another matter altogether, I suppose.
How can he eat massive amounts of pizza and loose weight?!
He probably forgot to mention that along with the message from monique@bigboobies.com, there was one from super_diet_pills@getyourmedsnow.com, offering revolutionary new pills (developed by doctors, no less, and reported in the New England Journal of Medicine) that would indeed allow him to eat pizza and lose weight.
I've often wondered if you have to eat pizza while you're taking these pills, and if so, what happens if you don't...
As for shared FAT32 drive, can't you mount the FAT32 in linux and symlink the mail folders directory in linux to the location on the windows
drive?
I had something like this working for a while with Netscape 4, when I was dual-booting between Red Hat 6 and Windows 98. As long as the mail program doesn't want to do any operations that aren't supported on FAT32 (I doubt it would), everything should be fine.
One thing that may trip you up is that the mail program creates index files for each of your mail folders to speed up access to them. In Netscape, at least, the naming convention was different in the two operating systems. That means that if you boot one OS, download some mail, and boot the other OS, (some of) your mail folders are newer than their indexes in that OS. This makes it think that you've hacked the folder manually, so it rebuilds the indexes to reflect your changes. Worse than that, neither version of the mail program recognises the other's index files, so it lists them as folders - but then claims they're not in the expected format.
My solution to this was rather than symlinking the directory that contained the mail folders, I created symlinks for each folder and its index file. This worked for me because the structure of my folders doesn't change very much. If you regularly create and rearrange folders, you might find it useful to run the Linux program with a wrapper script that recreates the symlinks.
Having said all of that, the Mozilla developers are probably aware of this problem. I wouldn't swear to it, but I think Mozilla uses the same naming convention for indexes on all platforms, so just symlink the directory and forget about it.
In the worst case, since reverse engineering the format would allow a person access to a copy protected data set, this would be a violation of the DMCA.
Yes, but... if you're trying to extract a document that you created, then presumably you own the copyright in it. The DMCA makes it illegal to bypass access controls only if you don't have the copyright holder's permission to do so. (I suppose Bill could sneak an "all your document are belong to us" clause into the EULA, but that would cause MS far worse problems than it could hope to solve.)
As for reverse engineering being a EULA violation... in the European Union (not sure about the USA), reverse engineering is legal if you need to do it to get hardware or software to interoperate with other hardware or software. That's exactly what you would be trying to do here. The fact that the EULA says you're not allowed to reverse engineer at all can't take away that exception.
Now, if you wrote a program to convert this new Word format to something else, and released it to the world, could MS sue you? Possibly for violating their EULA, possibly for infringing patents (if they managed to get a patent on part or all of the format), but perhaps not under the DMCA. Yes, you would be trafficking in a circumvention device, but that device has a substantial legitimate use - converting the user's own documents to some other format.
There are many more people producing Word documents than commercial CDs or DVDs. I think that, for most documents, the number of people who would want to read them is fairly small, regardless of whether they're in a DRM-encumbered format or not. It's therefore hard to claim that a great deal of financial harm results from your bypassing any access controls the document might have. Even if you do use this converter to read a document that you're not supposed to have access to, it would be up to the document creator to sue you, because it's their copyright you would be infringing.
If MS did decide to sue someone who had written such a converter, the case could be spun so as to generate so much bad publicity for them that regardless of the legal outcome, they would lose in the court of public opinion. But then, like you said, Bill isn't an idiot, so I doubt such a lawsuit would come to pass.
Thanks for that tip. I'll have to try it next time I get spam for a domain where the contact is a Hotmail or Yahoo account.
I still think, though, that if we deny DNS to the spammers, the more clueful of them will just bypass it altogether, using IP addresses directly. The upside there is that unless they own the address block, it's hard for them to forge the whois record, because it's not under their control.
Umm... what are you going to do when someone sends a bucketload of spam to accounts at the ISPs who participate in this scheme, proclaiming that you can buy penis extenders (or, for added irony, "prestigious degrees from non-accredited colleges";-) at http://udel.edu/~markpell/?
If any domain name can be deleted or redirected on the basis of a few emails (and remember we're dealing with people who are proven liars) then congratulations, you've just invented a way to DoS any site off the web. On the other hand, if you require some sort of proof that the website owner sent or approved of the spam, then the spammer can spam with impunity until that proof has been gathered.
Tangentially related to this, a lot of the spam I get promotes websites where the URL has just an IP address, bypassing DNS altogether. If your proposal takes off, I suspect many more spammers would do this. Probably, the reasons they do it are:
They don't want the overhead (in money or time) of registering a new domain, when the site will probably get pulled in a few weeks anyway.
Most web users aren't very aware of domain names (the fact that Slashdot interpolates them into our posts after links suggests this), so there's little advantage to the spammer having one. After all, how much brand recognition does your average spamhaus have? To put it another way, if they had brand recognition, they wouldn't need to spam.
The default behaviour for ping varies from one OS to another.
The Linux one (on the distros I've used, anyway) sends packets until you kill it, printing out the round trip time and TTL for each packet that comes back. It prints nothing for lost packets, except in the summary at the end. So yes, if you run the command on Linux, you've asked for 10 infinite loops.
The Solaris ping, though, sends just one packet, and prints "host is alive" or "no answer from host".
The Windows one (shudder) sends four packets and prints the round trip and TTL for all of them.
Having said all of that, I'm as puzzled as you are about what the grandparent post is trying to accomplish;-)
Well, here in Britain, "slag" means the waste from smelting the ore of a metal - much the same as in the US, I suppose. Strictly speaking, therefore, a melted hard drive platter isn't slag, but there's an analogy with something that's left over after intense heat.
The other meaning of "slag" in Britain is as a derogatory term for a debauched or promiscuous woman. Using it to someone's face is a good way to become acquainted with new forms of pain.;-)
You may have come across this suggestion already, so I don't know if it'll help, but anyway...
Mystery and crime writers often decide on the end first and then work backwards. I think you can apply this to all kinds of fiction. A story is like a journey, if only metaphorically, and you generally don't embark on a journey without having some idea of where you want to get to.
When I say "the end," I don't necessarily mean what happens in the last chapter, or the last paragraph. Rather, I mean what changes have happened by the time you get to the final page - what results does the story have? For example, if you're writing a story where the good guys face death at every turn, you should decide whether all of them will survive to the end of the book. If you decide to kill one of them off, that could happen in chapter 2, but it would still be one of the results of the story.
Once you've decided what your results will be, you can then decide on a plausible path (aka a plot) that allows all of them to happen. I've used this technique to restart a large fantasy novel which had been languishing for about 10 years while I found excuses to do other things. It's nearly doubled in size (from about 50,000 words to about 90,000), and I've been beavering away at it for some six months now, which is longer than I ever managed to stick at it before.
A site that you may find useful is run by the author Holly Lisle. She has an article describing this technique, plus lots of other good stuff.
It may be that both are right. 300,000 would probably be the number of clock ticks per second, while 1,000 would be the (average) number of instructions the machine performed per second. It's only fairly recently, with innovations like pipelining, that CPUs have been able to perform one instruction per clock tick.
A Z80 CPU, for example, which is of a similar vintage to your Apple ][, took at least 4 clock ticks to perform an instruction, while some took as many as 21. Executing an instruction requires several different parts of the CPU to operate on it in the right order. On a machine such as that, while a given part of the CPU is doing its job, the other parts have to sit idle, waiting for the result.
So, progress in hardware hasn't just been to make the computers smaller and run the clocks faster - it's also been to make the computer do more on each clock tick. If you could somehow get CSIRAC to run at 2GHz, or underclock a modern x86 to 300kHz (and you think Mozilla is slow now!), the modern machine would still wipe the floor with the old one.
...but then I do most of my surfing on a Sun Ultra 1 at work, so the first time I saw one of these banners, it was rather obvious that it wasn't a message from the operating system. The fact that the "dialogue box" scrolls when you move the page is a bit of a giveaway too.
I don't think these particular adverts are any more annoying than any other banners, but I haven't had to clean up after anyone who was suckered into installing whatever malware these guys were pushing. Even my girlfriend, who went through a phase of merrily installing anything she could get her hands on, never asked me why her computer was telling her that her 'net connection wasn't fast enough, and why the software that was supposed to fix it was plastered with adverts.
...but since you weren't very specific about subject matter, I might as well mention that I have a few dozen pictures (drawings and computer-generated) for perusal at my website. Start here.
I retain the copyright on these images, but I'm fairly permissive about what you can do with them - basically, whatever you like, as long as you don't sell them, or represent them as someone else's work. Enjoy!
Before posting a referral link, why not just have your software visit the referring site and detemine if it actually links to your page? This will defeat the referral advertisers.
Until the advertiser figures out that he can look at your referrer string (or put a ROT-13'd version of your URL into his referrer when he asks you for a page). When your script comes to check for a link back to you, he feeds your information into a CGI script that generates a page that includes the link you're looking for. Bingo! Your script is satisfied, and links to him.
Perhaps it won't get that far, but if it becomes a problem, bloggers will just have to check sites manually. It takes a pretty sophisticated script to work out that a hit from wilddonkeysex.com is probably bogus, but a human can reach that decision in about three seconds.
If there are too many to check manually, then perhaps they shouldn't backlink to a site until there have been more than a certain number of hits from it, or they should restrict their links page to the top 10, or the top 100.
I think you're right about link farms - it's quite possible for your site to link to too many others. Unless you organise the links in some way (which a script probably can't), it implies that you haven't thought about who's worth linking to and who isn't.
I run a website for a local club of amateur video-makers (stop sniggering at the back:-p ). We have rather a lot of links, but I've set up every one of them as the result of an email exchange. Either I asked the other website, or they asked me, if we would be interested in linking to each other. By this alone, our club is now in the top ten on a Google search for "camcorder". It's taken us two years to get there, but we can afford to be patient.
One is fantasy, for teenagers, and would be maybe 120 pages in paperback. The other is dystopian science fiction, for adults, and would be about 200 pages. Both are in plain HTML, without any kind of DRM. Older browsers might have trouble rendering the quotation marks and apostrophes properly, but Mozilla handles them fine, and that's [Ff]ree too.
Many people in this thread have grumbled that computer screens, whatever technology they may be based on, aren't suitable for reading large amounts of text. Here's a suggestion. Read as much of the book as you can bear on-screen, or as much as you need to decide whether it's worth finishing, whichever comes first. Then, if you think it's good enough, print it out and read the rest like that. Simple, eh? (It's so simple that I should patent it... no, let's not go there.)
OK, a printout isn't as nice to handle as a well-made paperback or hardback, but it's about as portable, it's a lot cheaper, you can print another if it gets lost or too dog-eared, and the words are the same in any case. Best of all, if you decide the book isn't worth reading, all it's cost you is the time you needed to reach that decision. (Yes, there's the cost of your net access while you download the book, and the running cost of your computer, but if money is tight enough that you have to worry about that, you probably shouldn't be wasting time reading novels;-)
You can find four (count 'em!) complete and unrestricted songs by yours truly here. They're in RealAudio, I'm afraid; at the time I didn't know any better. I'll get around to replacing them with MP3s or OGGs one of these years. I'm too old and cynical for a career as a musician now, so do whatever you like with them, as long as you don't sell them or represent them as someone else's work. Ta!
D'oh! You're right. Thanks for pointing that out. Five years is 3 1/3 doublings of CPU power, not a multiplication by 3 1/3. I think I got the longer term prediction right, though. In 30 years, we might (should?) have machines that are a billion times faster than today's, and RC5-64 might be crackable on a single machine in a year and a half. Give it 10 more years after that, they'll be a hundred times faster again, and a single machine might crack the cipher in a week.
I think my other point still stands, that ciphers will increase in complexity and key length, so that the ciphers of 30 or 40 years from now will take just as long to attack by brute force as today's do.
in several years time, the average laptop / home PC will be able to crank out the work that the distributed project did in a week or so...
I think you underestimate the scale of the problem - RSA's press release says there were over 300,000 people working on it for nearly five years. So, if Moore's Law continues to hold (a doubling of CPU power every 18 months or so), then in five years' time, computers will be, on average, 3 1/3 times faster than they are today. That means that you could repeat the RC5-64 "experiment" with 90,000 people, instead of 300,000, but it would still take nearly five years. Or you use the same number of people, and they'd be able to do it about 17 months.
I agree that, given enough doublings of CPU power, it will become feasible to crack RC5-64 with a single machine, but by my calculations, such a machine won't exist for 30 or 40 years. No doubt by then, if we're not already using quantum computers, we'll have something like RC20-65536, and cracking that will still need hundreds of thousands of machines to crunch numbers for years.
I imagine directors aren't happy about their movies being chopped up for TV and airline viewing, but because the TV station or airline pays a lot of money for the right to show the films, the directors are willing to cry all the way to the bank. There's probably also some discussion with the director as to what needs to be taken out for family viewing while retaining as much as possible of the "artistic vision". Clean Flicks don't seem to be doing either of those things.
One of the rights you have when you create something that's protected by copyright is the right to make or authorise "derivative works" - works that are recognisably based on what you've created. An example would be a movie version of a book. Another would be an edited version of a movie. Creating a derivative work without the permission of whoever owns the copyright in the original is illegal. End of story. I predict that this lawsuit will be thrown out before the lawyers on either side have warmed themselves up. If Clean Flicks had asked the copyright owners for permission to butcher their films, and (probably) paid a large wad of cash up front, they might have found themselves a comfortable niche in the market. As it is, I think they'll sink without trace, and good riddance to them.
Speaking as an artist myself (well, a writer, but it's the same difference), I can tell you that I'd be pretty narked if I found out that someone had taken one of my stories, removed all the swearing, sex and violence, and had published it as a "clean" version. If you can't handle the fact that the main pastime of one of my characters is the rape and torture of another character, go and read something else. Taking out that theme just to pander to someone's blinkered view of what art should be would remove the victim's main reason for doing what she does. When you finished the book, your reaction would not be "Oh, how sad," but "Huh? Why the <bleep> did she do that?"
Now, if you want to run my book through sed to remove all the swearing before you read it, that's your decision, and I accept there's nothing I can do to stop you. If you want to distribute a "patch" to my book, with instructions like "remove paragraph 523, delete the word 'burfle' from paragraph 524...", then this is a borderline case. The reader has to go get my original book, and can decide for themselves whether they want to read my version or yours. But if you go distributing your patched version, even if you make it clear that it's not my original, that's copyright infringement (unauthorised creation of a derivative work and unauthorised distribution of copyrighted works), and your ISP will get a take-down notice faster than you can say "DMCA sucks!"
Sorry if I seem to be ranting - flame and mod away - but I think this lawsuit should prove that sometimes, even a bad law can yield the right result.
My Windows 98 CD (first edition, I suppose you could call it) is also bootable. Booting from a CD requires that the BIOS support it. Older BIOSs don't do this, which might be why the earlier poster thought only Windows 2000 can do it.
You might also find that you need to tell the BIOS to check for a bootable CD at startup, as this option might not be set by default.
Er, there's a perfectly good control and alt within one-handed reach of the delete key.
There is now, yes. But twenty-or-so years ago, there wasn't. I forget the exact details, but one or the other of CTRL and ALT was only on the left-hand side of the keyboard, while DEL was only on the right. IBM probably picked CTRL-ALT-DEL because, with that layout, it was very hard to press that combination accidentally. Bear in mind that in those days, CTRL-ALT-DEL didn't bring up a dialogue box asking which program you wanted to kill. It rebooted straightaway, no questions asked, so pressing it accidentally wasn't something you could recover from.
...it's linked to in my sig. I'm not sure if it really qualifies as "dystopic," but we seem to be using a broad definition of the term here. The ending isn't exactly happy, but it is "as happy as possible." Think of the saying "if life hands you lemons, make lemonade." There are quite a few lemons handed out in this book;-)
Yes, the number of states would rapidly become unmanagable if you tried to hack together a finite state automaton to recognise a language such as A^nB^n for limited n. Actually, the problem isn't so much the number of states as the number of transitions between them, which is roughly the number of states multiplied by the number of symbols in your alphabet. This is sometimes known as the state-symbol product.
I suppose you could use some sort of regexp compiler to take the grunt work out of it. Ah... any programming language that has regexps has one of those built in anyway. Well, being able to write something like/A^nB^n/(n<=3) would be an advance on/^|AB|AABB|AAABBB$/, I suppose.
On the subject of education, I did a one-term course on theoretical computer science in general, and another specifically on formal languages. I can safely say they've been of no direct, practical use, but they provided a good foundation to what I studied in other courses and what I've learned since graduation.
Perhaps it's not important for everyone in computing to know what a DFA is, or understand the pumping lemma or the Church-Turing hypothesis, but it's necessary that some of us do. Otherwise, who will write the compilers for the next generation of programming languages? Who in the team that builds one of those compilers will tell the PHB that (the general case of) the halting problem is unsolvable, so that the marketing department really shouldn't be claiming that it will be able to detect infinite loops in a user's program before running them?
Slashdot Mirror
This is true, in that better spam filtering will allow the ISP to put off buying bigger disks for a while. But a crucial difference between disk space and CPU cycles is that you can reuse disk, but you can't reuse CPU.
From the ISP's point of view, it may not make much sense to spend, say, 5 microseconds to run a Bayesian filter on a 5K message, and then decide to dump it in /dev/null instead of /var/spool/mail, when half an hour later, the mailbox owner will read the subject line and delete the message anyway. Either way, the ISP has the 5K back, but the 5 microseconds are gone forever.
Of course, the mailbox owner may well take the view that he'd rather have the computer waste 5 microseconds deleting a message than him waste 5 whole seconds doing it himself. This is what angers me most about spam, more than the fact that it's driving up my costs: these bastards are stealing my time, and giving me nothing in return.
Absolutely. I'm getting near the point where I'd pay for something like that. But at the back of my mind is a little voice asking: "Why should you have to pay not to receive crap that has no benefit for you and that you never asked for in the first place?"Anything that improves the signal-to-noise ratio in peoples' inboxes is good, but I don't believe client-side filtering will make much of a dent in the spammers' market. The whole point of spam is that a spammer can afford to send a million messages and make maybe ten sales. Anybody who hates spam enough to install a client-side filter is probably already one of the 999,990 people who wouldn't have fallen for the scam even if they had read the message. But if the result is that a person's inbox becomes useable again, would they care how it happened?
Better yet, if this law really is as broad as claimed, most software for sending spam is now illegal in Michigan. It might even be illegal to operate an open relay, or to use one for spamming. Persuading law enforcement to go after a spammer on these grounds would be another matter altogether, I suppose.
How can he eat massive amounts of pizza and loose weight?!
He probably forgot to mention that along with the message from monique@bigboobies.com, there was one from super_diet_pills@getyourmedsnow.com, offering revolutionary new pills (developed by doctors, no less, and reported in the New England Journal of Medicine) that would indeed allow him to eat pizza and lose weight.
I've often wondered if you have to eat pizza while you're taking these pills, and if so, what happens if you don't...
I don't think I want an Inter- Continental Ballistic Missle aimed at my webpage, do you?
It happens already. It's called the Slashdot Effect ;-)
As for shared FAT32 drive, can't you mount the FAT32 in linux and symlink the mail folders directory in linux to the location on the windows drive?
I had something like this working for a while with Netscape 4, when I was dual-booting between Red Hat 6 and Windows 98. As long as the mail program doesn't want to do any operations that aren't supported on FAT32 (I doubt it would), everything should be fine.
One thing that may trip you up is that the mail program creates index files for each of your mail folders to speed up access to them. In Netscape, at least, the naming convention was different in the two operating systems. That means that if you boot one OS, download some mail, and boot the other OS, (some of) your mail folders are newer than their indexes in that OS. This makes it think that you've hacked the folder manually, so it rebuilds the indexes to reflect your changes. Worse than that, neither version of the mail program recognises the other's index files, so it lists them as folders - but then claims they're not in the expected format.
My solution to this was rather than symlinking the directory that contained the mail folders, I created symlinks for each folder and its index file. This worked for me because the structure of my folders doesn't change very much. If you regularly create and rearrange folders, you might find it useful to run the Linux program with a wrapper script that recreates the symlinks.
Having said all of that, the Mozilla developers are probably aware of this problem. I wouldn't swear to it, but I think Mozilla uses the same naming convention for indexes on all platforms, so just symlink the directory and forget about it.
In the worst case, since reverse engineering the format would allow a person access to a copy protected data set, this would be a violation of the DMCA.
Yes, but... if you're trying to extract a document that you created, then presumably you own the copyright in it. The DMCA makes it illegal to bypass access controls only if you don't have the copyright holder's permission to do so. (I suppose Bill could sneak an "all your document are belong to us" clause into the EULA, but that would cause MS far worse problems than it could hope to solve.)
As for reverse engineering being a EULA violation... in the European Union (not sure about the USA), reverse engineering is legal if you need to do it to get hardware or software to interoperate with other hardware or software. That's exactly what you would be trying to do here. The fact that the EULA says you're not allowed to reverse engineer at all can't take away that exception.
Now, if you wrote a program to convert this new Word format to something else, and released it to the world, could MS sue you? Possibly for violating their EULA, possibly for infringing patents (if they managed to get a patent on part or all of the format), but perhaps not under the DMCA. Yes, you would be trafficking in a circumvention device, but that device has a substantial legitimate use - converting the user's own documents to some other format.
There are many more people producing Word documents than commercial CDs or DVDs. I think that, for most documents, the number of people who would want to read them is fairly small, regardless of whether they're in a DRM-encumbered format or not. It's therefore hard to claim that a great deal of financial harm results from your bypassing any access controls the document might have. Even if you do use this converter to read a document that you're not supposed to have access to, it would be up to the document creator to sue you, because it's their copyright you would be infringing.
If MS did decide to sue someone who had written such a converter, the case could be spun so as to generate so much bad publicity for them that regardless of the legal outcome, they would lose in the court of public opinion. But then, like you said, Bill isn't an idiot, so I doubt such a lawsuit would come to pass.
Thanks for that tip. I'll have to try it next time I get spam for a domain where the contact is a Hotmail or Yahoo account.
I still think, though, that if we deny DNS to the spammers, the more clueful of them will just bypass it altogether, using IP addresses directly. The upside there is that unless they own the address block, it's hard for them to forge the whois record, because it's not under their control.
Umm... what are you going to do when someone sends a bucketload of spam to accounts at the ISPs who participate in this scheme, proclaiming that you can buy penis extenders (or, for added irony, "prestigious degrees from non-accredited colleges" ;-) at http://udel.edu/~markpell/?
If any domain name can be deleted or redirected on the basis of a few emails (and remember we're dealing with people who are proven liars) then congratulations, you've just invented a way to DoS any site off the web. On the other hand, if you require some sort of proof that the website owner sent or approved of the spam, then the spammer can spam with impunity until that proof has been gathered.
Tangentially related to this, a lot of the spam I get promotes websites where the URL has just an IP address, bypassing DNS altogether. If your proposal takes off, I suspect many more spammers would do this. Probably, the reasons they do it are:
The default behaviour for ping varies from one OS to another.
The Linux one (on the distros I've used, anyway) sends packets until you kill it, printing out the round trip time and TTL for each packet that comes back. It prints nothing for lost packets, except in the summary at the end. So yes, if you run the command on Linux, you've asked for 10 infinite loops.
The Solaris ping, though, sends just one packet, and prints "host is alive" or "no answer from host".
The Windows one (shudder) sends four packets and prints the round trip and TTL for all of them.
Having said all of that, I'm as puzzled as you are about what the grandparent post is trying to accomplish ;-)
The story was called Little Lost Robot. I don't know if it was in I, Robot, but it was in a later collection called The Complete Robot.
Well, here in Britain, "slag" means the waste from smelting the ore of a metal - much the same as in the US, I suppose. Strictly speaking, therefore, a melted hard drive platter isn't slag, but there's an analogy with something that's left over after intense heat.
The other meaning of "slag" in Britain is as a derogatory term for a debauched or promiscuous woman. Using it to someone's face is a good way to become acquainted with new forms of pain. ;-)
You may have come across this suggestion already, so I don't know if it'll help, but anyway...
Mystery and crime writers often decide on the end first and then work backwards. I think you can apply this to all kinds of fiction. A story is like a journey, if only metaphorically, and you generally don't embark on a journey without having some idea of where you want to get to.
When I say "the end," I don't necessarily mean what happens in the last chapter, or the last paragraph. Rather, I mean what changes have happened by the time you get to the final page - what results does the story have? For example, if you're writing a story where the good guys face death at every turn, you should decide whether all of them will survive to the end of the book. If you decide to kill one of them off, that could happen in chapter 2, but it would still be one of the results of the story.
Once you've decided what your results will be, you can then decide on a plausible path (aka a plot) that allows all of them to happen. I've used this technique to restart a large fantasy novel which had been languishing for about 10 years while I found excuses to do other things. It's nearly doubled in size (from about 50,000 words to about 90,000), and I've been beavering away at it for some six months now, which is longer than I ever managed to stick at it before.
A site that you may find useful is run by the author Holly Lisle. She has an article describing this technique, plus lots of other good stuff.
Good luck!
It may be that both are right. 300,000 would probably be the number of clock ticks per second, while 1,000 would be the (average) number of instructions the machine performed per second. It's only fairly recently, with innovations like pipelining, that CPUs have been able to perform one instruction per clock tick.
A Z80 CPU, for example, which is of a similar vintage to your Apple ][, took at least 4 clock ticks to perform an instruction, while some took as many as 21. Executing an instruction requires several different parts of the CPU to operate on it in the right order. On a machine such as that, while a given part of the CPU is doing its job, the other parts have to sit idle, waiting for the result.
So, progress in hardware hasn't just been to make the computers smaller and run the clocks faster - it's also been to make the computer do more on each clock tick. If you could somehow get CSIRAC to run at 2GHz, or underclock a modern x86 to 300kHz (and you think Mozilla is slow now!), the modern machine would still wipe the floor with the old one.
...but then I do most of my surfing on a Sun Ultra 1 at work, so the first time I saw one of these banners, it was rather obvious that it wasn't a message from the operating system. The fact that the "dialogue box" scrolls when you move the page is a bit of a giveaway too.
I don't think these particular adverts are any more annoying than any other banners, but I haven't had to clean up after anyone who was suckered into installing whatever malware these guys were pushing. Even my girlfriend, who went through a phase of merrily installing anything she could get her hands on, never asked me why her computer was telling her that her 'net connection wasn't fast enough, and why the software that was supposed to fix it was plastered with adverts.
...but since you weren't very specific about subject matter, I might as well mention that I have a few dozen pictures (drawings and computer-generated) for perusal at my website. Start here.
I retain the copyright on these images, but I'm fairly permissive about what you can do with them - basically, whatever you like, as long as you don't sell them, or represent them as someone else's work. Enjoy!
Until the advertiser figures out that he can look at your referrer string (or put a ROT-13'd version of your URL into his referrer when he asks you for a page). When your script comes to check for a link back to you, he feeds your information into a CGI script that generates a page that includes the link you're looking for. Bingo! Your script is satisfied, and links to him.
Perhaps it won't get that far, but if it becomes a problem, bloggers will just have to check sites manually. It takes a pretty sophisticated script to work out that a hit from wilddonkeysex.com is probably bogus, but a human can reach that decision in about three seconds.
If there are too many to check manually, then perhaps they shouldn't backlink to a site until there have been more than a certain number of hits from it, or they should restrict their links page to the top 10, or the top 100.
I think you're right about link farms - it's quite possible for your site to link to too many others. Unless you organise the links in some way (which a script probably can't), it implies that you haven't thought about who's worth linking to and who isn't.
I run a website for a local club of amateur video-makers (stop sniggering at the back :-p ). We have rather a lot of links, but I've set up every one of them as the result of an email exchange. Either I asked the other website, or they asked me, if we would be interested in linking to each other. By this alone, our club is now in the top ten on a Google search for "camcorder". It's taken us two years to get there, but we can afford to be patient.
...follow the link in my .sig to read them.
One is fantasy, for teenagers, and would be maybe 120 pages in paperback. The other is dystopian science fiction, for adults, and would be about 200 pages. Both are in plain HTML, without any kind of DRM. Older browsers might have trouble rendering the quotation marks and apostrophes properly, but Mozilla handles them fine, and that's [Ff]ree too.
Many people in this thread have grumbled that computer screens, whatever technology they may be based on, aren't suitable for reading large amounts of text. Here's a suggestion. Read as much of the book as you can bear on-screen, or as much as you need to decide whether it's worth finishing, whichever comes first. Then, if you think it's good enough, print it out and read the rest like that. Simple, eh? (It's so simple that I should patent it... no, let's not go there.)
OK, a printout isn't as nice to handle as a well-made paperback or hardback, but it's about as portable, it's a lot cheaper, you can print another if it gets lost or too dog-eared, and the words are the same in any case. Best of all, if you decide the book isn't worth reading, all it's cost you is the time you needed to reach that decision. (Yes, there's the cost of your net access while you download the book, and the running cost of your computer, but if money is tight enough that you have to worry about that, you probably shouldn't be wasting time reading novels ;-)
You can find four (count 'em!) complete and unrestricted songs by yours truly here. They're in RealAudio, I'm afraid; at the time I didn't know any better. I'll get around to replacing them with MP3s or OGGs one of these years. I'm too old and cynical for a career as a musician now, so do whatever you like with them, as long as you don't sell them or represent them as someone else's work. Ta!
D'oh! You're right. Thanks for pointing that out. Five years is 3 1/3 doublings of CPU power, not a multiplication by 3 1/3. I think I got the longer term prediction right, though. In 30 years, we might (should?) have machines that are a billion times faster than today's, and RC5-64 might be crackable on a single machine in a year and a half. Give it 10 more years after that, they'll be a hundred times faster again, and a single machine might crack the cipher in a week.
I think my other point still stands, that ciphers will increase in complexity and key length, so that the ciphers of 30 or 40 years from now will take just as long to attack by brute force as today's do.
I think you underestimate the scale of the problem - RSA's press release says there were over 300,000 people working on it for nearly five years. So, if Moore's Law continues to hold (a doubling of CPU power every 18 months or so), then in five years' time, computers will be, on average, 3 1/3 times faster than they are today. That means that you could repeat the RC5-64 "experiment" with 90,000 people, instead of 300,000, but it would still take nearly five years. Or you use the same number of people, and they'd be able to do it about 17 months.
I agree that, given enough doublings of CPU power, it will become feasible to crack RC5-64 with a single machine, but by my calculations, such a machine won't exist for 30 or 40 years. No doubt by then, if we're not already using quantum computers, we'll have something like RC20-65536, and cracking that will still need hundreds of thousands of machines to crunch numbers for years.
...it usually does.
I imagine directors aren't happy about their movies being chopped up for TV and airline viewing, but because the TV station or airline pays a lot of money for the right to show the films, the directors are willing to cry all the way to the bank. There's probably also some discussion with the director as to what needs to be taken out for family viewing while retaining as much as possible of the "artistic vision". Clean Flicks don't seem to be doing either of those things.
One of the rights you have when you create something that's protected by copyright is the right to make or authorise "derivative works" - works that are recognisably based on what you've created. An example would be a movie version of a book. Another would be an edited version of a movie. Creating a derivative work without the permission of whoever owns the copyright in the original is illegal. End of story. I predict that this lawsuit will be thrown out before the lawyers on either side have warmed themselves up. If Clean Flicks had asked the copyright owners for permission to butcher their films, and (probably) paid a large wad of cash up front, they might have found themselves a comfortable niche in the market. As it is, I think they'll sink without trace, and good riddance to them.
Speaking as an artist myself (well, a writer, but it's the same difference), I can tell you that I'd be pretty narked if I found out that someone had taken one of my stories, removed all the swearing, sex and violence, and had published it as a "clean" version. If you can't handle the fact that the main pastime of one of my characters is the rape and torture of another character, go and read something else. Taking out that theme just to pander to someone's blinkered view of what art should be would remove the victim's main reason for doing what she does. When you finished the book, your reaction would not be "Oh, how sad," but "Huh? Why the <bleep> did she do that?"
Now, if you want to run my book through sed to remove all the swearing before you read it, that's your decision, and I accept there's nothing I can do to stop you. If you want to distribute a "patch" to my book, with instructions like "remove paragraph 523, delete the word 'burfle' from paragraph 524...", then this is a borderline case. The reader has to go get my original book, and can decide for themselves whether they want to read my version or yours. But if you go distributing your patched version, even if you make it clear that it's not my original, that's copyright infringement (unauthorised creation of a derivative work and unauthorised distribution of copyrighted works), and your ISP will get a take-down notice faster than you can say "DMCA sucks!"
Sorry if I seem to be ranting - flame and mod away - but I think this lawsuit should prove that sometimes, even a bad law can yield the right result.
My Windows 98 CD (first edition, I suppose you could call it) is also bootable. Booting from a CD requires that the BIOS support it. Older BIOSs don't do this, which might be why the earlier poster thought only Windows 2000 can do it.
You might also find that you need to tell the BIOS to check for a bootable CD at startup, as this option might not be set by default.
There is now, yes. But twenty-or-so years ago, there wasn't. I forget the exact details, but one or the other of CTRL and ALT was only on the left-hand side of the keyboard, while DEL was only on the right. IBM probably picked CTRL-ALT-DEL because, with that layout, it was very hard to press that combination accidentally. Bear in mind that in those days, CTRL-ALT-DEL didn't bring up a dialogue box asking which program you wanted to kill. It rebooted straightaway, no questions asked, so pressing it accidentally wasn't something you could recover from.
Yes, the number of states would rapidly become unmanagable if you tried to hack together a finite state automaton to recognise a language such as A^nB^n for limited n. Actually, the problem isn't so much the number of states as the number of transitions between them, which is roughly the number of states multiplied by the number of symbols in your alphabet. This is sometimes known as the state-symbol product.
I suppose you could use some sort of regexp compiler to take the grunt work out of it. Ah... any programming language that has regexps has one of those built in anyway. Well, being able to write something like /A^nB^n/(n<=3) would be an advance on /^|AB|AABB|AAABBB$/, I suppose.
On the subject of education, I did a one-term course on theoretical computer science in general, and another specifically on formal languages. I can safely say they've been of no direct, practical use, but they provided a good foundation to what I studied in other courses and what I've learned since graduation.
Perhaps it's not important for everyone in computing to know what a DFA is, or understand the pumping lemma or the Church-Turing hypothesis, but it's necessary that some of us do. Otherwise, who will write the compilers for the next generation of programming languages? Who in the team that builds one of those compilers will tell the PHB that (the general case of) the halting problem is unsolvable, so that the marketing department really shouldn't be claiming that it will be able to detect infinite loops in a user's program before running them?