I compete regularly at TopCoder and some of my friends joke that I'm "punching the monkey". Of course, if so, I've punched the little bugger over 400 times now....
I once played netquake for 18 hours. A couple bathroom breaks, and I may have switched servers a few times too. I regularly used to play the entire map cycle at least once in a sitting on my favorite CTF server, which typically took at least three hours.
A lot of this was during the summer of 1997, IIRC. I was on a 36/12 schedule and my average for the entire summer was probably in excess of 8 hours per 24 of Quake. It was actually a great experience. I got very good at the game, and really enjoyed 1v1 when I started playing that, even though I didn't know any locals to practice with (at my level). I also formed a clan and wrote a mod for netquake, and found a bunch of bugs in the binary and in other popular mods. And I did get a lot of other things done, since the Quake was mostly during time I was working as a labmonkey anyway. It's a lot easier to step out of a big game of Quake to answer a users question than it would be to leave behind a game of Starcraft, for example....
How does high value make a PDA inadequate? Just treat the PDA like something valuable, as valuable as the contracts it can potentially sign. Put it in a safe when you aren't using it, for example. Were you going to sign things more valuable than a PDA's weight of gold? Probably, OK then, a PDA's weight of cashier's checks with lots of zeroes on them? Exactly.
And if you need multiple parties to sign a document before it is valid, you can either just put multiple locks on the safe, or use multiple safes and use an appropriate multiparty signature protocol.
If you are only trying to make it possible for one person to digitally sign documents with their own key, it can be much simpler than all that. Just write a module for a PDA that generates the key internally and can sign documents on it, and wave lots of warning signs at the user when they do something that would copy their private key off the PDA. If you never run the PDA software on anything you don't read first (or put any untrusted software on it), how can you screw up? Obviously you need a PDA where the data transfer can be adminstered from the PDA side, not the random-untrusted-PC side, but the software work for this seems like a lot less than custom-tailoring and auditing an entire linux kernel. You could even physically mangle the communication link so that it works in one direction only, and when you sign something, manually transcribe the result, which should be a reasonably short hex string. Or only sign hashes of documents (which is typical anyway) and also input the hash by hand, but then you have to trust the computer generating the hash, since you don't get to inspect the plaintext on the PDA as you sign it.
What are you concerned about Tempest radiation for, anyway? Maybe the system bus would leak information about the private key, but the _monitor_? All it should be doing is displaying the contract, and the contract doesn't need to be secret... indeed, it will not remain so if there is ever a dispute about the signers.
Be careful pointing the finger about ignorant mistakes.
TSP cannot be worse than NP-complete, because it is obviously in NP. Phrased as a decision problem (is there a Hamilton path through this graph shorter than length y?) it is trivial to verify a solution in polynomial time. If you can verify in P, you can solve in NP.
Note that rephrasing as a decision problem doesn't change the order much, because you can just do a binary search with O(log N) steps where each is a decision subproblem. Also note that transforming it into a decision problem is *necessary* to discuss its NP-completeness, because the very concept is only defined for decision problems.
I, too, have friends who have worked loading packages for UPS. The volume of packages they handle is too great to bother with most of what you describe. (Most of them worked night shift, yes.) One of them told me once about accidentally putting his foot through a package and damaging it, and I asked him how often that happened. He said less than once a night, out of the thousands of items he handles. I heard a lot more horror stories about people shipping stuff like diving ballast than anything about destroying packages, deliberately or not.
Anyway, the damage done to the packages shown in the article was obviously caused by more than a couple energetic kicks from a loader.
In my opinion, what you are observing is the problems TCP already has handling congestion. If opening parallel connections increases your throughput, this just shows that TCP was not correctly utilizing the available bandwidth in the first place.
If the Slashdot community wants any recognition of their views, the most effective way would be to form a PAC, accept donations (to cover travel and administrative costs, and possibly to influence some elections via advertising), elect volunteers to coordinate for particular activities or areas of Federal law, tally people's opinions on various legislative proposals, and send physical representatives (lobbyists) knocking on Congresspersons' doors.
Or, if that seems like too much trouble, even though it is how most effective political expression occurs in the US, then at the very least join organizations that already reflect your views. For example, people who are for free trade in crypto and against software patents should become paying ACM members, since the ACM already lobbies on behalf of those views.
Yes, I think the fact that any of this is necessary to get our own government ("of the people") to respect our views is terribly sad and reeks of corruption. But we live in the real world, and this is how things are done. It is the rare, determined, charismatic person who can make any progress changing our government outside this corrupt system, and most of those people are (sadly) ultimately assimilated into it.
have you *tried* it? i think they are just being conservative about quality, not saying its a stable release version until they are satisfied. it already sounds better at 128k than any mp3 codec ive tried (recent lame and all flavors of fraunhofer included).
I think 95%+ of the people playing, for example, counter-strike nowadays have never even *seen* a real bot. I get accused of cheating on a pretty regular basis, and anyone who thinks I am a bot has never seen a netquake bot that could fire rockets out of its ass with perfect predictive accuracy.
Hmm... I don't know exactly why metadata operations should be slow on XFS, but my understanding is that XFS is primarily useful for:
Journaling, which depending on how it is implemented, would make exactly when metadata writes occur a lot less important. So you may just be seeing the forced sync of 5-second-old dirty metadata from bdflush fouling up the performance. Perhaps XFS was designed with a much looser disk/memory binding in mind.
Very high throughput sequential I/O on very large files on very large disks/RAIDs (think video editing), which might involve optimizations or tradeoffs that are not synergistic with low-latency metadata operations.
Little time or patience required, I can explain the delete behavior easily. (Not that I would label myself an FS hacker... distributed, maybe.)
Modern inode-based filesystems scatter both files and metadata across the disk, to reduce average seek time and fragmentation problems (free blocks are always nearby when using cylinder groups, unless the disk is very very full). So deletion can involve visiting many areas of the disk, and also traversing the allocation tree for each file and adding its blocks back to the free list (or vector).
MS-DOS FAT filesystems have most of the metadata tightly clustered in the two file allocation tables... only subdirectories containing filenames, some metadata, and pointers into the FAT are stored outside. (32 bytes each.) So for FAT you have to seek, read, and write once (usually) for each subdirectory, but for all the normal files you just make one pass on each FAT, which probably involves no seeking, since they fit entirely within a single cylinder on modern disks.
DOS just made this process seem excruciatingly slow because you were typically using it without caching and on floppies. FAT-based filesystems are a bit of a naive approach, and their average-case performance, especially under load, can be quite poor if you can't cache the entire FAT in memory. But they aren't all that awful.
And Diablo 1 was notorious for cheating! The correct answer is to centralize the important things on the server, because otherwise clients are free to modify them. Checksums are just obfuscation, unless you do crypto things to make a private key necessary to log in to your account... which, IMO, is overkill for a game. Of course, you also have to avoid bugs in your database system that completely circumvent the usual login procedure:)
Your post shows a complete lack of awareness of the history here--I think you are trolling.
Police believe criminals talk to each other by e-mail to avoid phone taps.
So, won't criminals just use PGP and anonymous remailers (based in other countries with strict privacy laws) now?
Citizens should be pushing for constitutional protection of privacy (in nations that don't have it already; it exists in the US, but only by Supreme Court fiat, AFAICT). Unless you think it's OK to read your mail before even charging you with anything.
Using RF radiation to trap ionized gas or plasma at a surface is very useful in space. You can maintain a small barrier in which solid particles can be vaporized harmlessly before they interact with the hull proper. You could also collect fuel and mineral resources, although you might want a larger scoop than your hull for that.
What can this provide that you wouldn't get from a full implementation of Windows services in Wine, or running DOS apps via DOSEMU or VMWare?
Aside from the sarcastic remarks about instability, I can only imagine one thing--use of Windows device drivers. Seems like reverse-engineering and rewriting specific drivers would be less effort.
Man pages are actually very good IMO. Sometimes it is tough to find the right man page, if you don't know the command or library call you need to do something, but once you find it, the information you need is almost always there.
Try what Loki is using: SDL. It handles video, audio, and input, and has ports for Linux, Win32, MacOS, and BeOS.
As for only loading the parts you need to run... ahh, the wonders of demand paging. That is the job of the OS nowadays.
The only real problems with writing Linux games, IMO, are immature support for hardware-accelerated OpenGL, and lack of library support for weird media types.
Your suggested uses sound very much like the Langston field from _The Mote in God's Eye_. The first seems like a very bright idea, and it might apply in other situations where finished surfaces are exposed to high-velocity debris at relatively low pressures... self-cleaning rooms or fan blades with integrated ozonators would be cool:)
But I don't follow you on the second. [Hot] fusion already involves plasma and ways of handling plasma without touching it... how exactly does adding another layer of *cold* plasma help?
Or no electrodes at all... a lot of fluorescent tubes just require an radio wave of appropriate frequency and magnitude... hold one near a power line some time.
The point here is *low energy* plasma, and no need for containment... neon in tubes is typically energized to about 10^4 C, and breaking the tube puts an end to the fun.
The article has a rather high hype-to-explanation ratio... so I went looking for the patent they mention. Not so technical as to be incomprehensible, and more useful than the article, IMO. My first impression is that the advance here is the impedance matching system used to maintain the plasma, which allows the compact equipment and low power requirements. (I suppose it searches for the natural resonance of whatever ions you have between the electrodes... just find the lowest-energy state/standing wave at which it remains permittive/permeable.) And if you have low power plus no sealed chamber (1 atm, random molecular gases allowed), it pretty much follows that the result is a low temperature plasma, since plasma tends to radiate continuously. The scalability aspect is nice too... good for more than a toy.
Either I'm misunderstanding something, or the sterilization is done by the radiation from the plasma... basically just using the plasma as an efficient UV lamp. (The sterilization patent talks about sterilizing liquids or gels up to 2cm deep... I can't see doing that with the surface interactions, which might be sufficient for polished tools and the like.) This doesn't appear to be sufficient for a low-volume irradiation system for food, which is unfortunate, because I like my hamburgers juicy. Oh well. (Of course, prions survive irradiation anyway, so I would still have to worry about BSE.)
There is also a separate patent for "surface shielding". Might be fun to set up on your car:) I'm having trouble figuring out how the leakage from this system would be less detectable than the reflected radio waves it would disperse, though... I suppose if you do it right, all the radiation is absorbed by the gases you are ionizing?
AIM and ICQ work fine... I'm not getting charged for them... there are clients for lots of platforms... so who cares about making a standard to interoperate with them? Only people that see a "market" and wish they were part of it.
I really can't bring myself to care until the technology improves. When someone invents the fully distributed datagram-only version with automatic encryption and message signing and no central point of failure, I will be the first to sign up. (I've been muttering about doing that to IRC for years, I suppose I should sit down and code something--except I hate writing GUIs, and that is the part most people would notice.) Until then, unless AOL is doing something weird with my messages, or tries to extract a fee from me, what difference does it make that Micros~1 and whoever else can't join the party?
As long as we are going to make a "standard", lets redo the protocol from the ground up and fix all the flaws... and yes, I've seen the IETF draft standard, I think its fugly. Publishing standards for the existing tools would be nice, but what right has anyone to force AOL to open their protocols? Personally I don't see anything sinister in AOL acquiring Mirabilis and Nullsoft and Netscape... they just want to make sure those tools continue to exist, if only so that the value of the Internet experience of their own customers is improved.
It's a little harder than the article says, actually. The proxy has to grok the map, too, or you fire at walls with people behind them.
I wrote a Q1 proxy... but it was designed to be run on the SERVER. (I had to do magical loopback things to convince a Q1 server that each connection was from a different client, because it mostly ignored port numbers... required multihoming on most IP stacks.) It was meant to fix some lame problems with existing CTF mods (they couldn't even reliably kick players or enforce team colors), and to help players make macros to report their status to teammates. I was even working on some weird features like a good spectator mode (unlimited connections, ghosting another player without help from the server), and server side demos (merging all the event streams from different clients to make a single demo that included all the gamplay).
Anyway, I spent a lot more time on that, and tools to query lists of servers (before gamespy existed), than I ever did on cheats. But Q1 wasn't very popular anymore by the time I started on the more interesting stuff, so I never really finished it due to lack of demand:(
Slashdot Mirror
I compete regularly at TopCoder and some of my friends joke that I'm "punching the monkey". Of course, if so, I've punched the little bugger over 400 times now....
I once played netquake for 18 hours. A couple bathroom breaks, and I may have switched servers a few times too. I regularly used to play the entire map cycle at least once in a sitting on my favorite CTF server, which typically took at least three hours.
A lot of this was during the summer of 1997, IIRC. I was on a 36/12 schedule and my average for the entire summer was probably in excess of 8 hours per 24 of Quake. It was actually a great experience. I got very good at the game, and really enjoyed 1v1 when I started playing that, even though I didn't know any locals to practice with (at my level). I also formed a clan and wrote a mod for netquake, and found a bunch of bugs in the binary and in other popular mods. And I did get a lot of other things done, since the Quake was mostly during time I was working as a labmonkey anyway. It's a lot easier to step out of a big game of Quake to answer a users question than it would be to leave behind a game of Starcraft, for example....
How does high value make a PDA inadequate? Just treat the PDA like something valuable, as valuable as the contracts it can potentially sign. Put it in a safe when you aren't using it, for example. Were you going to sign things more valuable than a PDA's weight of gold? Probably, OK then, a PDA's weight of cashier's checks with lots of zeroes on them? Exactly.
And if you need multiple parties to sign a document before it is valid, you can either just put multiple locks on the safe, or use multiple safes and use an appropriate multiparty signature protocol.
Schneier actually suggests using a handheld in the article.
At least that means this solution is obvious. Generate your keypair on your PDA, and then secure it physically.
If you are only trying to make it possible for one person to digitally sign documents with their own key, it can be much simpler than all that. Just write a module for a PDA that generates the key internally and can sign documents on it, and wave lots of warning signs at the user when they do something that would copy their private key off the PDA. If you never run the PDA software on anything you don't read first (or put any untrusted software on it), how can you screw up? Obviously you need a PDA where the data transfer can be adminstered from the PDA side, not the random-untrusted-PC side, but the software work for this seems like a lot less than custom-tailoring and auditing an entire linux kernel. You could even physically mangle the communication link so that it works in one direction only, and when you sign something, manually transcribe the result, which should be a reasonably short hex string. Or only sign hashes of documents (which is typical anyway) and also input the hash by hand, but then you have to trust the computer generating the hash, since you don't get to inspect the plaintext on the PDA as you sign it.
What are you concerned about Tempest radiation for, anyway? Maybe the system bus would leak information about the private key, but the _monitor_? All it should be doing is displaying the contract, and the contract doesn't need to be secret... indeed, it will not remain so if there is ever a dispute about the signers.
Be careful pointing the finger about ignorant mistakes.
TSP cannot be worse than NP-complete, because it is obviously in NP. Phrased as a decision problem (is there a Hamilton path through this graph shorter than length y?) it is trivial to verify a solution in polynomial time. If you can verify in P, you can solve in NP.
Note that rephrasing as a decision problem doesn't change the order much, because you can just do a binary search with O(log N) steps where each is a decision subproblem. Also note that transforming it into a decision problem is *necessary* to discuss its NP-completeness, because the very concept is only defined for decision problems.
I, too, have friends who have worked loading packages for UPS. The volume of packages they handle is too great to bother with most of what you describe. (Most of them worked night shift, yes.) One of them told me once about accidentally putting his foot through a package and damaging it, and I asked him how often that happened. He said less than once a night, out of the thousands of items he handles. I heard a lot more horror stories about people shipping stuff like diving ballast than anything about destroying packages, deliberately or not.
Anyway, the damage done to the packages shown in the article was obviously caused by more than a couple energetic kicks from a loader.
In my opinion, what you are observing is the problems TCP already has handling congestion. If opening parallel connections increases your throughput, this just shows that TCP was not correctly utilizing the available bandwidth in the first place.
If the Slashdot community wants any recognition of their views, the most effective way would be to form a PAC, accept donations (to cover travel and administrative costs, and possibly to influence some elections via advertising), elect volunteers to coordinate for particular activities or areas of Federal law, tally people's opinions on various legislative proposals, and send physical representatives (lobbyists) knocking on Congresspersons' doors.
Or, if that seems like too much trouble, even though it is how most effective political expression occurs in the US, then at the very least join organizations that already reflect your views. For example, people who are for free trade in crypto and against software patents should become paying ACM members, since the ACM already lobbies on behalf of those views.
Yes, I think the fact that any of this is necessary to get our own government ("of the people") to respect our views is terribly sad and reeks of corruption. But we live in the real world, and this is how things are done. It is the rare, determined, charismatic person who can make any progress changing our government outside this corrupt system, and most of those people are (sadly) ultimately assimilated into it.
have you *tried* it? i think they are just being conservative about quality, not saying its a stable release version until they are satisfied. it already sounds better at 128k than any mp3 codec ive tried (recent lame and all flavors of fraunhofer included).
I think 95%+ of the people playing, for example, counter-strike nowadays have never even *seen* a real bot. I get accused of cheating on a pretty regular basis, and anyone who thinks I am a bot has never seen a netquake bot that could fire rockets out of its ass with perfect predictive accuracy.
Hmm... I don't know exactly why metadata operations should be slow on XFS, but my understanding is that XFS is primarily useful for:
Just my thoughts....
Little time or patience required, I can explain the delete behavior easily. (Not that I would label myself an FS hacker... distributed, maybe.)
Modern inode-based filesystems scatter both files and metadata across the disk, to reduce average seek time and fragmentation problems (free blocks are always nearby when using cylinder groups, unless the disk is very very full). So deletion can involve visiting many areas of the disk, and also traversing the allocation tree for each file and adding its blocks back to the free list (or vector).
MS-DOS FAT filesystems have most of the metadata tightly clustered in the two file allocation tables... only subdirectories containing filenames, some metadata, and pointers into the FAT are stored outside. (32 bytes each.) So for FAT you have to seek, read, and write once (usually) for each subdirectory, but for all the normal files you just make one pass on each FAT, which probably involves no seeking, since they fit entirely within a single cylinder on modern disks.
DOS just made this process seem excruciatingly slow because you were typically using it without caching and on floppies. FAT-based filesystems are a bit of a naive approach, and their average-case performance, especially under load, can be quite poor if you can't cache the entire FAT in memory. But they aren't all that awful.
And Diablo 1 was notorious for cheating! The correct answer is to centralize the important things on the server, because otherwise clients are free to modify them. Checksums are just obfuscation, unless you do crypto things to make a private key necessary to log in to your account... which, IMO, is overkill for a game. Of course, you also have to avoid bugs in your database system that completely circumvent the usual login procedure :)
Your post shows a complete lack of awareness of the history here--I think you are trolling.
I thought the little network programmers were inside my TV... now you say they are in my VCR, too?
From the article:
So, won't criminals just use PGP and anonymous remailers (based in other countries with strict privacy laws) now?
Citizens should be pushing for constitutional protection of privacy (in nations that don't have it already; it exists in the US, but only by Supreme Court fiat, AFAICT). Unless you think it's OK to read your mail before even charging you with anything.
I see that you are stuck in a 1 atm world :)
Using RF radiation to trap ionized gas or plasma at a surface is very useful in space. You can maintain a small barrier in which solid particles can be vaporized harmlessly before they interact with the hull proper. You could also collect fuel and mineral resources, although you might want a larger scoop than your hull for that.
What can this provide that you wouldn't get from a full implementation of Windows services in Wine, or running DOS apps via DOSEMU or VMWare?
Aside from the sarcastic remarks about instability, I can only imagine one thing--use of Windows device drivers. Seems like reverse-engineering and rewriting specific drivers would be less effort.
Man pages are actually very good IMO. Sometimes it is tough to find the right man page, if you don't know the command or library call you need to do something, but once you find it, the information you need is almost always there.
Try what Loki is using: SDL. It handles video, audio, and input, and has ports for Linux, Win32, MacOS, and BeOS.
As for only loading the parts you need to run... ahh, the wonders of demand paging. That is the job of the OS nowadays.
The only real problems with writing Linux games, IMO, are immature support for hardware-accelerated OpenGL, and lack of library support for weird media types.
Your suggested uses sound very much like the Langston field from _The Mote in God's Eye_. The first seems like a very bright idea, and it might apply in other situations where finished surfaces are exposed to high-velocity debris at relatively low pressures... self-cleaning rooms or fan blades with integrated ozonators would be cool :)
But I don't follow you on the second. [Hot] fusion already involves plasma and ways of handling plasma without touching it... how exactly does adding another layer of *cold* plasma help?
Or no electrodes at all... a lot of fluorescent tubes just require an radio wave of appropriate frequency and magnitude... hold one near a power line some time.
The point here is *low energy* plasma, and no need for containment... neon in tubes is typically energized to about 10^4 C, and breaking the tube puts an end to the fun.
The article has a rather high hype-to-explanation ratio... so I went looking for the patent they mention. Not so technical as to be incomprehensible, and more useful than the article, IMO. My first impression is that the advance here is the impedance matching system used to maintain the plasma, which allows the compact equipment and low power requirements. (I suppose it searches for the natural resonance of whatever ions you have between the electrodes... just find the lowest-energy state/standing wave at which it remains permittive/permeable.) And if you have low power plus no sealed chamber (1 atm, random molecular gases allowed), it pretty much follows that the result is a low temperature plasma, since plasma tends to radiate continuously. The scalability aspect is nice too... good for more than a toy.
Either I'm misunderstanding something, or the sterilization is done by the radiation from the plasma... basically just using the plasma as an efficient UV lamp. (The sterilization patent talks about sterilizing liquids or gels up to 2cm deep... I can't see doing that with the surface interactions, which might be sufficient for polished tools and the like.) This doesn't appear to be sufficient for a low-volume irradiation system for food, which is unfortunate, because I like my hamburgers juicy. Oh well. (Of course, prions survive irradiation anyway, so I would still have to worry about BSE.)
There is also a separate patent for "surface shielding". Might be fun to set up on your car :) I'm having trouble figuring out how the leakage from this system would be less detectable than the reflected radio waves it would disperse, though... I suppose if you do it right, all the radiation is absorbed by the gases you are ionizing?
AIM and ICQ work fine... I'm not getting charged for them... there are clients for lots of platforms... so who cares about making a standard to interoperate with them? Only people that see a "market" and wish they were part of it.
I really can't bring myself to care until the technology improves. When someone invents the fully distributed datagram-only version with automatic encryption and message signing and no central point of failure, I will be the first to sign up. (I've been muttering about doing that to IRC for years, I suppose I should sit down and code something--except I hate writing GUIs, and that is the part most people would notice.) Until then, unless AOL is doing something weird with my messages, or tries to extract a fee from me, what difference does it make that Micros~1 and whoever else can't join the party?
As long as we are going to make a "standard", lets redo the protocol from the ground up and fix all the flaws... and yes, I've seen the IETF draft standard, I think its fugly. Publishing standards for the existing tools would be nice, but what right has anyone to force AOL to open their protocols? Personally I don't see anything sinister in AOL acquiring Mirabilis and Nullsoft and Netscape... they just want to make sure those tools continue to exist, if only so that the value of the Internet experience of their own customers is improved.
Or does someone have evidence to the contrary?
It's a little harder than the article says, actually. The proxy has to grok the map, too, or you fire at walls with people behind them.
I wrote a Q1 proxy... but it was designed to be run on the SERVER. (I had to do magical loopback things to convince a Q1 server that each connection was from a different client, because it mostly ignored port numbers... required multihoming on most IP stacks.) It was meant to fix some lame problems with existing CTF mods (they couldn't even reliably kick players or enforce team colors), and to help players make macros to report their status to teammates. I was even working on some weird features like a good spectator mode (unlimited connections, ghosting another player without help from the server), and server side demos (merging all the event streams from different clients to make a single demo that included all the gamplay).
Anyway, I spent a lot more time on that, and tools to query lists of servers (before gamespy existed), than I ever did on cheats. But Q1 wasn't very popular anymore by the time I started on the more interesting stuff, so I never really finished it due to lack of demand :(