Copyright worked as long as it was not rigidly enforfced. You know, that whole "no part may be reproduced in any way" thing... academia has survived throughout the years by MASS coppying - I myself, as a high school student, even, regularly recieve hundreds of copyrighted coppied pages. Education has, until very recently, been protected under the fair use conditions set by courts and somewhat by laws. And it *NEEDS* that freedom in order to work. Books go out of print that teachers need, etc. Education cannot work as an industry - it's TEACHING, not MANUFACTURING. Though I admit the line gets blurred with some institutions, sadly.
Reguardless, no, Copyright didn't work in entirety. And the Bono extensions just make the damn thing fall apart around itself by making it a laughable law. LIMITED TIMES (20 years from publication, tough shit after that, or something like that) are what we need.
Humans don't live forever. Corporations (which *are not people*, damnitall... somebody get the SC to reverse that one... ^_~), which have tremendously longer potential lifetimes, should NOT get the benefits of being a citizen - lobbying power, etc. There can be, of course, copyright laws that apply to corporations, but they should not exceed and preferably be SHORTER than times for individuals (thus giving an economic *advantage* of letting the artists own their own works - it's the companie's [indirectly] longer).
Grrrr. Sorry for this and all other posts I write. I'm just angry at a system that I see as crumbling around me that I wish were better.
No, they don't *have* to connect to the network. Using HAM radio data pathways, 802.11[ab], *gasp* modems, etc., it's entirely possible to build a second network.
Though I do have one question, and I really really want an answer. How do we prevent IP-space collisions in a fully p2p system? Or should we just use IPv6, ARP, and some method of collision resolution (and if so, how do we handle collision resolution? ^_^).
The single fact is this: almost all costs associated with upstream bandwidth I can think of are one-time or constant-rate costs: laying fiber and buying fiber equipment (once), paying for electricity bill for router (constant). Downstream bandwidth will, of course, vary depending on implementation, BUT - it seems that it'd be also one-time and constant-rate costs too... installing the link, powering the equipment to keep it linked. There's some maintainence costs, too, but those should roughly scale with the user base - it's easy for one tech to maintain lots of downstream connections, and there's a much much smaller number of upstream connections, and most problems with that, if you set it up right, are a call to the upstream provider away from being fixed.
The backbone providers have a rather special situation. They may not necessarily make a profit from the network, but it is still in their best interest to keep the network going! Imagine the bad press... "AT&T shut down most of its backbone today..."
To follow up on your clever routing things, I've long thought that a modified gnutella connect protocol could flatten the network down to match the lower router base of the internet - exactly what you said - so all the nodes on the inside of, say, RR's routers, know each other and don't ever cross RR's uplinks, except for maybe a few (or at worst maybe ~a connection per servent, as opposed to ~six) connections which do have to cross it. Maybe even a modified version of that hypercube idea that got posted a while back - run that inside RR's network and use the boring gnutella protocol to channel out to the outside world? Maybe, maybe not.
Anyway, this irks me. Who wants to start more freenets or Internet3 (^^) - I'd be more than willing to contribute whatever I can.
I'm building the sweetest desktop workstation I can afford to pay for ASAP then not buying another system until *all* this crap blows over or runs itself into the ground. As I'm considering a ham licence already, I'll at least have a potential freenet there, and with the development of 802.11 and other protocols, I'm sure I can get others. I'll pay for internet access if I can aford it and iff it *stays a flat price*. This talk of per-bit pricing scares me. (it costs no more to keep a fiber line lit if I use it or no. And the number of lines that need to be lit scales with the number of users paying monthly fees.) I know there's no free lunch, but there's already a bandwidth cap (of 31 days * 80KiBps) on my DSL line.
But I digress. The thing that scares me *more* than the loss of the internet (since we can just throw up freenets - HAM, 802.11, etc, as I mentioned) is the loss of control over the software in my computer. Screw DRM - it's not touching my new system. A plain old boring ACPI BIOS, DRM-free HDDs, cracked DVD drive, and a good CDRW drive... everything runs just fine under Linux, with open source and open program channels. Sounds good to me. ^^
Though, while I'm rambling, what we need as geeks wanting a freenet is a grand unified peer-to-peer system. If possible (I'm not sure it is, but, this is just a sketchy list to try to convey the ideas I'm talking about) the system should:
0) Be open. GPL/BSD dual licence, or somesuch. 1) Be truly peer to peer. The network can elect transient ultrapeers (ala Gnutella v0.6) should nodes "wish." 2) Allow for file sharing (duh), ideally with some way of tracking files (unique IDs on the network... maybe just a combination of name, md5, and size information?) so they can be referenced exactly by other documents (equivelant of web, right there). It must still be searchable like gnutella or FastTrack or whatever. 3) Allow for distributed computing with fully authenticated cores. Something like the (now defunct and abandoned?) distributed.net v3, a proposal I liked very much. Ideally we'd find a way of allowing the dist-comp aspect to scale from low-bandwidth dist-comp to ultra-high bandwidth requirements so that the same API, maybe not actual implementation, could be used universally. 3.1) Allow dist-comp cores to come in a variety of languages. A java core would be great for the truly paranoid, whereas a highly tweaked assembler version would be great for speed. Allow the owner of the computers to decide which cores and whom to trust.
Anybody got any pointers? ^^
Ok, I'm sorry, I got a little off there. I tend to ramble, what can I say.
Automagic updates are all well and good, as long as there's good authentication, preferably good encryption, and at least some amount of "Hey, User, you want to install this?" with the default being [Yes], not no, and of course a pointer to more information.
Brilliant here has (apparently?) done away with all three. They just do it (like Nike), and from the sound of the article, they are not even very secure about the way they do it.
The reassuring thing (for the moment) is that so far these tactics of behind-the-scenes trojans have been confined to leaf nodes - to my knowledge, no routers etc. have had this kind of shit happen to them. As long as the major routing backbones of the internet never become 0wned, there's a modicum of hope for restoring order to the network (banning IPs at the fringes of the backbones until they shape up?) should an emergency occur (banning IPs always scared me, so I don't necessarily like that solution, but it's the easiest and the one that jumped to mind first. I'm sure people more clever than I can think of better ones).
OTOH, 1M fringe nodes can, as the article says, be unstopable. If somebody were truly evil and wrote a decentralized worm (never called home, only talked with other copies of itself), it would be incredibly hard to stop such a beast, and the DDOS commands could be given in an anonymous, untrackable way (can anybody imagine the worms playing Dining Cryptographers? ^_^) [Dining Cryptographers would be anonymous as long as the line wasn't tapped. And I'm sure with some good encryption over the links, it'd be anonymous for all practical purposes anyway.]
Y'know, as bad as it'd be, I'd want to see such a worm (just it's source, I *swear* - I'm not about to go risking the internet's well-being - you have to admit it'd be an interesting read). Maybe the vx community has something similar as a proof of concept?
Hey - given the sheep mentality of 90+% of the people participating in this economy, it could be a good thing if the news media become optimisitic without proof. IANA Economist, but if people start spending, it seems that that could trigger an upwards feedback loop (optimism --> spending --> profits --> optimism, repeat) to help get us out of this "recession."
Though you're right - the companies have to be careful. Rasing the prices, no matter how much optimism, might cause people to second guess purchacing decisions (works in any industry, not just computers) and thus hold us in the recession.
Ok, I'm going to stop looking into my crystal ball now. It's prolly defective anyway.
Ick. Ick ick ickie ick. PIM and EMAIL integration? Urhm... I don't like it in Outlook (I don't *want* my mother again in software) and I don't think I'd like it (though I suppose I could be convinced) if it were repeated.
Integrated PGP support - cool! Just don't re-implement the PGP protocol in the email app. Call out to other programs. UNIX makes that easy enough, and I'd presume Windows would too.
Split-view Inbox? Uh, why?! Are multiple folders *that* hard to deal with? Besides, a split inbox would just waste screen realestate (and thus be in line with MS products like VisualStudio).
Built in instant messanging is along the same lines. Two (or three) applications can't be *that* hard to do, people. Between the introduction of rich text emails and these people's designs, it seems like the email client is destined to become the UberApplication, eventually folding into itself all other programs. Ack!
SPAM auto-reporting. Looks good on paper, probably going to be much more a pain than it's worth. Most spammers fake their headers or other nasty tricks. [We should just legalize thurough beatings of any spammers we manage to ID in person. ^_- ]
Mouseover contact information. Ok, cool - easy to do, not much bloat. This actually belongs in an email client. Though contact information should be stored, of course, in an LDAP server.
Smart email notification - been there, done that.
All powerful right clicking. Seems to just be the usual objection to the quirkiness of MS's UIs. Seems like a good thing for all programs - sensible interface design.
Easy access to message templets - same as above - just a cleaner interface, please!
IIRC, Mann had, at least for a while, used his equipment to rotate the world 90-degrees for his vision. He also had some form of reality mediator program running for a while, according to his website - that could be any one of a number of things, including assisted memory.
The sudden switch of the world back 90 degrees and stripping of assisted memory would really be a traumatic experience for the brain.
AFAIK, there is no known formula for the n-th decimal digit of pi. However, there is one for the n-th base sixteenth (hexadecimal) digit.
D = (1/16^n)[(1/(2n+1)) + (1/(2n+3)) + [two other similar terms]].
I'm not sure the first two terms are right, but the formula has a (1/16^n) term multiplied by four 1/(an+b) terms, where a and b are constants and n is the n'th digit of Pi.
Do these people just not get it? I could cite the VCR, the audio-cassette... etc. Technology has historically *eventually* *always* won (cf. Marxism). Now, I know it's teleologic to attempt to apply that to the future, but it really makes sense in this case - the DMCA did not prevent DeCSS, instead only made it a bigger problem because people mirrored it who would otherwise not have touched it.
I could also cite their MULTI-BILLION DOLLAR PROFITS. So what, exactly, is the problem? Not taking every cent out of the consumers... hmm.
Geez... anybody have a REALLY REALLY LARGE ClueStick?
And hey, maybe the Senators will realize that *they* wouldn't want such devices as the SSSCA-all-over-again will require.
Hello, this is a helpless citizen of Planet Earth watching the system fall into a mess.
For starters, these "productivity" measures never really make anybody more productive - they just dash morale and send people into ruts of drudgery and just going through the motions of work. Our bodies and brains are just not ABLE to do one thing for hours upon hours on end - we find several ways around that, be it day-dreaming, checking/., reading / writing email, etc.
Secondly, cowering behind "Email viruses" is just FUCKING DUMB. If a car manufacturer (here goes the analogy, sorry) were selling cars that had easily broken engines (such that a bump in the road would cause them to break - think any one of the bugs in Outlook. "begin" jumps to mind, along with all the virus-carrying security holes), no company that depended on cars would order their workers not to drive! But since it's The Microsoft Way, it can't be Microsoft's or Outlook's fault, no.... it must be.... uh.... those NAUGHTY JOKES employess FREELY CIRCULATE for the boring purpose of TAKING THEIR MIND OFF WORK FOR TEN SECONDS and relaxing. JEEZUS. (Sorry for shouting. I get angry.)
And "content filtering systems" - uh... how many times are we going to be shown that these are TOTALLY ineffective in any meaninful sense of the word "Effective." (Other than to make their parent companies $$.) They ban more than they let through, even of innocuous stuff.
Grrrrrrrrr. Grunk. Groo.::hopes that some day all will be well and the world will be corporate-idiocy free::
Formally, a n-CPU system is really equivelant just a Turing Machine with n times more internal information (data memory, data cache, disk data, etc. may all be considered "internal state", with the program residing on the tape still). Thus, if one UTM couldn't ever solve the halting problem, N of them in parallel could never solve it either.
Now, I can't speak towards a NDUTM, or an array of NDUTMs or QC. Would be nice (though unlikely) if every problem we can ask had a P-time solution (string theory computation?) - but the truth probably is that there's P, NP, NP-complete, and generally unsolvable problems (halting). But computation theory is still new, so maybe what I consider "probable" isn't.
Ok, wahoo, I just (theoretically) coppied *everything* about the GF4 card, against all current IP laws, and now have several DVDs full of stuff. Let's even say I got the software to make sense of it, and that my systems at home can render chip designs fast.
Now, what do I do?
Well... do I want a GF4 card? Yes. Ok, option: hand data over to IC mfmr, PCB mfmr + stuffer, wait for them to process my qty 1 order, pay upwards of 3x what it cost to by an NVIDIA or other implementation GF4... no advantage over just buying one off the shelf.
Do I want to steal NVIDIA's design and make hella money? Yes - money is good, right? Well... Ok, I hand over the chip data to the IC mfmr, I get a multilayer PCB drilling press, tracer / etchant station, etc, etc... and start churning them out. Oh... sale. Right - Have to say I have them. Make www.immitationgf4.com, offer them for sale. NVIDIA gets wind of what I'm doing and simply issues a statement that "immitationgf4.com" stole our data and is producing so-called GF4s without verification, etc - we do not support these boards. Gamers, OEMs, everybody serious and even most not-serious people flee my site, don't buy, and I'm out the cost of some very very expensive toys.
Now, if a major company steals NVIDIA's data... again, NVIDIA just says "Look what they did! We don't support those cards, they are likely broken, etc." And their sales don't drop much.
Yes, the code may be running on my machine, but given that I'm *required* to have a net connection the entire time, it seems a little too risky - how do I know that the only thing going across the wire is the app? How do I know that my data isn't being sent back? And most importantly - can I save to my local system and not some ASP's computers. The ASPs may say that it allieviates the need for backups, but all it really does is take total control of your data from your hands and places it in somebody else's.
The White Man's Burden (the poem), I've been told, was probably intended as satire. I don't remember it well enough to quote citations, but that's what my teacher said. Take it with a grain of salt, obviously.
Close enough? Sure, for government work, maybe (=)) but it's not journaled and as such has NO guarantee of meta-data or data consistancy, as opposed to JFS, XFS, EXT3, JFFS, etc.
What needs to be done is the creation of a universal virtual architecture for the whole damn thing. I don't want to manage ten programs on twenty computers. I want there to be one program on the twenty computers that check in with another one to not only get work units / assignments but the cores too. So I can say to my controller box, computer 1 can run projects A,B,D,F,and Q, with weighted priorities of.1,.2,.4,.2, and.1 respectively. If no work is avalable, fall to next highest priority. Then have computer 2 work on A,C,E,and Q, randomly pick between A,C,E and only run Q if none of those projects have outstanding work (works for A,C,E being things like MOSIX).
The controller should also know the locations of the projects masters, so it acts as a data proxy unless otherwise allowed and can fetch newer versions of the cores in real time. Of course, the cores can just languish on disk until I sign them, should I chose to.
IOW, something like the now-abandoned distributed.net v3.
No, fat32 is not journaled. Beowulf_boy just thinks it is because Windows tends not to keep lots and lots of dirty data in RAM (as opposed to *nix, etc.) and so seems to have better crash-safety.
It's a lot like the mac's FS - HFS isn't journaled either, but since MacOS n (n 10.0) had single threading disk access and didn't keep a lot of dirty data in ram, the disk was usually in sync.
So... if a blue laser (l = 440nm) has a better sampling resolution than a red one (l = 700nm), why couldn't a blue laser device read a disk burned by a red laser? Or would we have to get to UV lasers and use 350nm light?
TIA. _Knots
Re:THE BIG FREAKING POINT.
on
SSSCA Hearing
·
· Score: 1
TCP session hashes may be a clever solution to the problem as it now stands, but there's a so-called wheat-and-chaff secrecy system (don't have links, but google should be able to find something for the more curious) that is [almost?] completely undetectable to a listening party.
You have a document you want to send to me, but you worry that somebody's watching our TCP packets. Or worse - your document has the same MD5 hash as one of the multi-million songs out there by the RIAA (c'mon, there had to be an address-space collision sooner or later). We agree, using a key generation algorithm (ala SSH and others), to encrypt everything. Ok, cool - our packets now look like random noise addressed to each other. Now we make it an order of magnitude worse. We agree on a second key that is not used for encryption. We take this key, maybe do a transform on it, and then just use it as a flow-controlling shift register. If the current bit is one, we transmit a valid packet of data. If the current bit is zero, we transmit a valid packet that has part of/dev/[u]random for its payload! Shift the key, and repeat. If the key runs out before transmission's completed, we can then agree on another key.
The power of this method is that, as long as the second key (well, really, shared-secret) generation is secure, the whole stream is nonsense. Any attempt to session-hash it is meaningless, and any except the hardest attempts to decode it will be difficult.
The method works better with smaller packets, obviously - smaller amount of data == more nondetermanism (from the point of view of one not in possession of the shared secret).
So maybe secrecy is our last and final option. And hey! If the hardware lockdown is really *that* anal, it would detect a PS/2 keyboard logger as a means of circumvention [can't have somebody capturing somebody else's passphrases]! =)
RMS has his own motives, which you would better be asking him about. Ditto for everybody in the world, no?
As for my own motives, I see OSS as the best model to produce software - right in front of everybody, so the whole world can potentially learn, criticize (constructivly, ideally), or ignore. People certainly have the freedom to produce closed-source systems, BUT they do not have the right to do half of what they do with copyright these days (shrink-wrap EULAs) or force me to run proprietary software (gotta love open standards). That's just my view - proprietary software's cool; what I'm opposed to are abuses of copyright (preventing RE, the DMCA, the SSSCA, etc) as applied to software and generally crappy or sub-standard proprietary software (In my view, that includes Windows and virtually every piece of software to come from Microsoft, but not exclusive to MS).
If I see Windows running somewhere, I might laugh and tease about them using a joke OS, but hey, if it works for what they need it to, I'm not going to try to *make* them run Linux or BSD or MacOS or whatever - and people tease me for running Linux ('s fine), but they don't try to make me run Windows [corporations do, though, which is totally uncool].
If an arbitrary company doesn't want to use OSS, I'm not about to try to force them. Now, if it's the company I work at - remember, they pay me for my knowledge in this area, *I* am going to want to use Linux, but that's because I'm most comfortable with it. If they tell me I have to use Windows, I'm going to go down kicking and screaming. But a for-sale UNIX.... well, I probably could manage that, if they insisted.
Remember, OSS is a *gift* to the world - the world can choose to snub it if it so desires. There will always be somebody with the proverbial itch to scratch. I for one, will continue using Linux (and helping its development in my own small way).
Now, one area where this gets grey in my mind is companies like RedHat. They didn't develop the software, they just package it with pretty colors and sell support. So one could perhaps argue that they are just support contractors that *happen* to sell you physical media containing a free product already available online. (Hey, if that line of logic works, maybe we won't have to worry about companies violating the GPL - because suddenly they'd be responsible for *everything* in a code-tree that they sold.) But now what about cheapbytes and such? I suppose they're just selling physical media containing stuff already available on their FTP site [and really, they do just charge for media and shipping and little else, especially not the software] - so I guess that's not a problem.
Oh no... between the DMCA and whoever owns the genes for human eyeballs, fingers, etc.... I'm going to be paying A TON to keep living. Though of course I'll be getting my Cease-and-Desist letters from the gene companies in jail because of the DMCA.
And I'll be stuck there until the media corps and the gene corps sue each other out of existence, fighting over my body parts' legality. Though they'll both agree that I can't have them.
The DMCA effectively outlaws higher math. Think about it - data encryption is just complex higher math. Reversing it requires some skill with a data logger / oscilloscope, etc., and quite a bit of work, these days, shuttling mathematics around on paper. At least as far as I've seen.... maybe I'm wrong.
All somebody has to do to ban Calculus class forever from all highschools in the US is make some encryption based on integration or differentiation. I'm sure that's already happened.... I can see it now: "Calculus, the study of Differential Equations, and all of number theory have been declared unlawful because their primary purpose has become the circumvention of encryption."
[Moderators - it's supposed to be a joke. Mod appropriately]
Slashdot Mirror
Copyright worked as long as it was not rigidly enforfced. You know, that whole "no part may be reproduced in any way" thing... academia has survived throughout the years by MASS coppying - I myself, as a high school student, even, regularly recieve hundreds of copyrighted coppied pages. Education has, until very recently, been protected under the fair use conditions set by courts and somewhat by laws. And it *NEEDS* that freedom in order to work. Books go out of print that teachers need, etc. Education cannot work as an industry - it's TEACHING, not MANUFACTURING. Though I admit the line gets blurred with some institutions, sadly.
Reguardless, no, Copyright didn't work in entirety. And the Bono extensions just make the damn thing fall apart around itself by making it a laughable law. LIMITED TIMES (20 years from publication, tough shit after that, or something like that) are what we need.
Humans don't live forever. Corporations (which *are not people*, damnitall... somebody get the SC to reverse that one... ^_~), which have tremendously longer potential lifetimes, should NOT get the benefits of being a citizen - lobbying power, etc. There can be, of course, copyright laws that apply to corporations, but they should not exceed and preferably be SHORTER than times for individuals (thus giving an economic *advantage* of letting the artists own their own works - it's the companie's [indirectly] longer).
Grrrr. Sorry for this and all other posts I write. I'm just angry at a system that I see as crumbling around me that I wish were better.
-Knots
No, they don't *have* to connect to the network. Using HAM radio data pathways, 802.11[ab], *gasp* modems, etc., it's entirely possible to build a second network.
Though I do have one question, and I really really want an answer. How do we prevent IP-space collisions in a fully p2p system? Or should we just use IPv6, ARP, and some method of collision resolution (and if so, how do we handle collision resolution? ^_^).
-Knots
Thank you. I wish I had modpoints.
The single fact is this: almost all costs associated with upstream bandwidth I can think of are one-time or constant-rate costs: laying fiber and buying fiber equipment (once), paying for electricity bill for router (constant). Downstream bandwidth will, of course, vary depending on implementation, BUT - it seems that it'd be also one-time and constant-rate costs too... installing the link, powering the equipment to keep it linked. There's some maintainence costs, too, but those should roughly scale with the user base - it's easy for one tech to maintain lots of downstream connections, and there's a much much smaller number of upstream connections, and most problems with that, if you set it up right, are a call to the upstream provider away from being fixed.
The backbone providers have a rather special situation. They may not necessarily make a profit from the network, but it is still in their best interest to keep the network going! Imagine the bad press... "AT&T shut down most of its backbone today..."
To follow up on your clever routing things, I've long thought that a modified gnutella connect protocol could flatten the network down to match the lower router base of the internet - exactly what you said - so all the nodes on the inside of, say, RR's routers, know each other and don't ever cross RR's uplinks, except for maybe a few (or at worst maybe ~a connection per servent, as opposed to ~six) connections which do have to cross it. Maybe even a modified version of that hypercube idea that got posted a while back - run that inside RR's network and use the boring gnutella protocol to channel out to the outside world? Maybe, maybe not.
Anyway, this irks me. Who wants to start more freenets or Internet3 (^^) - I'd be more than willing to contribute whatever I can.
-knots
I'm building the sweetest desktop workstation I can afford to pay for ASAP then not buying another system until *all* this crap blows over or runs itself into the ground. As I'm considering a ham licence already, I'll at least have a potential freenet there, and with the development of 802.11 and other protocols, I'm sure I can get others. I'll pay for internet access if I can aford it and iff it *stays a flat price*. This talk of per-bit pricing scares me. (it costs no more to keep a fiber line lit if I use it or no. And the number of lines that need to be lit scales with the number of users paying monthly fees.) I know there's no free lunch, but there's already a bandwidth cap (of 31 days * 80KiBps) on my DSL line.
But I digress. The thing that scares me *more* than the loss of the internet (since we can just throw up freenets - HAM, 802.11, etc, as I mentioned) is the loss of control over the software in my computer. Screw DRM - it's not touching my new system. A plain old boring ACPI BIOS, DRM-free HDDs, cracked DVD drive, and a good CDRW drive... everything runs just fine under Linux, with open source and open program channels. Sounds good to me. ^^
Though, while I'm rambling, what we need as geeks wanting a freenet is a grand unified peer-to-peer system. If possible (I'm not sure it is, but, this is just a sketchy list to try to convey the ideas I'm talking about) the system should:
0) Be open. GPL/BSD dual licence, or somesuch.
1) Be truly peer to peer. The network can elect transient ultrapeers (ala Gnutella v0.6) should nodes "wish."
2) Allow for file sharing (duh), ideally with some way of tracking files (unique IDs on the network... maybe just a combination of name, md5, and size information?) so they can be referenced exactly by other documents (equivelant of web, right there). It must still be searchable like gnutella or FastTrack or whatever.
3) Allow for distributed computing with fully authenticated cores. Something like the (now defunct and abandoned?) distributed.net v3, a proposal I liked very much. Ideally we'd find a way of allowing the dist-comp aspect to scale from low-bandwidth dist-comp to ultra-high bandwidth requirements so that the same API, maybe not actual implementation, could be used universally.
3.1) Allow dist-comp cores to come in a variety of languages. A java core would be great for the truly paranoid, whereas a highly tweaked assembler version would be great for speed. Allow the owner of the computers to decide which cores and whom to trust.
Anybody got any pointers? ^^
Ok, I'm sorry, I got a little off there. I tend to ramble, what can I say.
-knots
Automagic updates are all well and good, as long as there's good authentication, preferably good encryption, and at least some amount of "Hey, User, you want to install this?" with the default being [Yes], not no, and of course a pointer to more information.
Brilliant here has (apparently?) done away with all three. They just do it (like Nike), and from the sound of the article, they are not even very secure about the way they do it.
The reassuring thing (for the moment) is that so far these tactics of behind-the-scenes trojans have been confined to leaf nodes - to my knowledge, no routers etc. have had this kind of shit happen to them. As long as the major routing backbones of the internet never become 0wned, there's a modicum of hope for restoring order to the network (banning IPs at the fringes of the backbones until they shape up?) should an emergency occur (banning IPs always scared me, so I don't necessarily like that solution, but it's the easiest and the one that jumped to mind first. I'm sure people more clever than I can think of better ones).
OTOH, 1M fringe nodes can, as the article says, be unstopable. If somebody were truly evil and wrote a decentralized worm (never called home, only talked with other copies of itself), it would be incredibly hard to stop such a beast, and the DDOS commands could be given in an anonymous, untrackable way (can anybody imagine the worms playing Dining Cryptographers? ^_^) [Dining Cryptographers would be anonymous as long as the line wasn't tapped. And I'm sure with some good encryption over the links, it'd be anonymous for all practical purposes anyway.]
Y'know, as bad as it'd be, I'd want to see such a worm (just it's source, I *swear* - I'm not about to go risking the internet's well-being - you have to admit it'd be an interesting read). Maybe the vx community has something similar as a proof of concept?
-Knots
Hey - given the sheep mentality of 90+% of the people participating in this economy, it could be a good thing if the news media become optimisitic without proof. IANA Economist, but if people start spending, it seems that that could trigger an upwards feedback loop (optimism --> spending --> profits --> optimism, repeat) to help get us out of this "recession."
Though you're right - the companies have to be careful. Rasing the prices, no matter how much optimism, might cause people to second guess purchacing decisions (works in any industry, not just computers) and thus hold us in the recession.
Ok, I'm going to stop looking into my crystal ball now. It's prolly defective anyway.
-Knots
Ick. Ick ick ickie ick. PIM and EMAIL integration? Urhm... I don't like it in Outlook (I don't *want* my mother again in software) and I don't think I'd like it (though I suppose I could be convinced) if it were repeated.
Integrated PGP support - cool! Just don't re-implement the PGP protocol in the email app. Call out to other programs. UNIX makes that easy enough, and I'd presume Windows would too.
Split-view Inbox? Uh, why?! Are multiple folders *that* hard to deal with? Besides, a split inbox would just waste screen realestate (and thus be in line with MS products like VisualStudio).
Built in instant messanging is along the same lines. Two (or three) applications can't be *that* hard to do, people. Between the introduction of rich text emails and these people's designs, it seems like the email client is destined to become the UberApplication, eventually folding into itself all other programs. Ack!
SPAM auto-reporting. Looks good on paper, probably going to be much more a pain than it's worth. Most spammers fake their headers or other nasty tricks. [We should just legalize thurough beatings of any spammers we manage to ID in person. ^_- ]
Mouseover contact information. Ok, cool - easy to do, not much bloat. This actually belongs in an email client. Though contact information should be stored, of course, in an LDAP server.
Smart email notification - been there, done that.
All powerful right clicking. Seems to just be the usual objection to the quirkiness of MS's UIs. Seems like a good thing for all programs - sensible interface design.
Easy access to message templets - same as above - just a cleaner interface, please!
Ah, hell, that was unnecessary.
-Knots
IIRC, Mann had, at least for a while, used his equipment to rotate the world 90-degrees for his vision. He also had some form of reality mediator program running for a while, according to his website - that could be any one of a number of things, including assisted memory.
The sudden switch of the world back 90 degrees and stripping of assisted memory would really be a traumatic experience for the brain.
_knots
AFAIK, there is no known formula for the n-th decimal digit of pi. However, there is one for the n-th base sixteenth (hexadecimal) digit.
D = (1/16^n)[(1/(2n+1)) + (1/(2n+3)) + [two other similar terms]].
I'm not sure the first two terms are right, but the formula has a (1/16^n) term multiplied by four 1/(an+b) terms, where a and b are constants and n is the n'th digit of Pi.
-knots
There goes the neighborhood.
Do these people just not get it? I could cite the VCR, the audio-cassette... etc. Technology has historically *eventually* *always* won (cf. Marxism). Now, I know it's teleologic to attempt to apply that to the future, but it really makes sense in this case - the DMCA did not prevent DeCSS, instead only made it a bigger problem because people mirrored it who would otherwise not have touched it.
I could also cite their MULTI-BILLION DOLLAR PROFITS. So what, exactly, is the problem? Not taking every cent out of the consumers... hmm.
Geez... anybody have a REALLY REALLY LARGE ClueStick?
And hey, maybe the Senators will realize that *they* wouldn't want such devices as the SSSCA-all-over-again will require.
_knots
Hello, this is a helpless citizen of Planet Earth watching the system fall into a mess.
/., reading / writing email, etc.
::hopes that some day all will be well and the world will be corporate-idiocy free::
For starters, these "productivity" measures never really make anybody more productive - they just dash morale and send people into ruts of drudgery and just going through the motions of work. Our bodies and brains are just not ABLE to do one thing for hours upon hours on end - we find several ways around that, be it day-dreaming, checking
Secondly, cowering behind "Email viruses" is just FUCKING DUMB. If a car manufacturer (here goes the analogy, sorry) were selling cars that had easily broken engines (such that a bump in the road would cause them to break - think any one of the bugs in Outlook. "begin" jumps to mind, along with all the virus-carrying security holes), no company that depended on cars would order their workers not to drive! But since it's The Microsoft Way, it can't be Microsoft's or Outlook's fault, no.... it must be.... uh.... those NAUGHTY JOKES employess FREELY CIRCULATE for the boring purpose of TAKING THEIR MIND OFF WORK FOR TEN SECONDS and relaxing. JEEZUS. (Sorry for shouting. I get angry.)
And "content filtering systems" - uh... how many times are we going to be shown that these are TOTALLY ineffective in any meaninful sense of the word "Effective." (Other than to make their parent companies $$.) They ban more than they let through, even of innocuous stuff.
Grrrrrrrrr. Grunk. Groo.
_knots
Incorrect.
Formally, a n-CPU system is really equivelant just a Turing Machine with n times more internal information (data memory, data cache, disk data, etc. may all be considered "internal state", with the program residing on the tape still). Thus, if one UTM couldn't ever solve the halting problem, N of them in parallel could never solve it either.
Now, I can't speak towards a NDUTM, or an array of NDUTMs or QC. Would be nice (though unlikely) if every problem we can ask had a P-time solution (string theory computation?) - but the truth probably is that there's P, NP, NP-complete, and generally unsolvable problems (halting). But computation theory is still new, so maybe what I consider "probable" isn't.
_knots
Ok, wahoo, I just (theoretically) coppied *everything* about the GF4 card, against all current IP laws, and now have several DVDs full of stuff. Let's even say I got the software to make sense of it, and that my systems at home can render chip designs fast.
Now, what do I do?
Well... do I want a GF4 card? Yes. Ok, option: hand data over to IC mfmr, PCB mfmr + stuffer, wait for them to process my qty 1 order, pay upwards of 3x what it cost to by an NVIDIA or other implementation GF4... no advantage over just buying one off the shelf.
Do I want to steal NVIDIA's design and make hella money? Yes - money is good, right? Well... Ok, I hand over the chip data to the IC mfmr, I get a multilayer PCB drilling press, tracer / etchant station, etc, etc... and start churning them out. Oh... sale. Right - Have to say I have them. Make www.immitationgf4.com, offer them for sale. NVIDIA gets wind of what I'm doing and simply issues a statement that "immitationgf4.com" stole our data and is producing so-called GF4s without verification, etc - we do not support these boards. Gamers, OEMs, everybody serious and even most not-serious people flee my site, don't buy, and I'm out the cost of some very very expensive toys.
Now, if a major company steals NVIDIA's data... again, NVIDIA just says "Look what they did! We don't support those cards, they are likely broken, etc." And their sales don't drop much.
No?
_knots
Internet applications bug me.
Yes, the code may be running on my machine, but given that I'm *required* to have a net connection the entire time, it seems a little too risky - how do I know that the only thing going across the wire is the app? How do I know that my data isn't being sent back? And most importantly - can I save to my local system and not some ASP's computers. The ASPs may say that it allieviates the need for backups, but all it really does is take total control of your data from your hands and places it in somebody else's.
_knots
The White Man's Burden (the poem), I've been told, was probably intended as satire. I don't remember it well enough to quote citations, but that's what my teacher said. Take it with a grain of salt, obviously.
(Note to self: reread.)
_knots
Close enough? Sure, for government work, maybe (=)) but it's not journaled and as such has NO guarantee of meta-data or data consistancy, as opposed to JFS, XFS, EXT3, JFFS, etc.
_Knots
I sugguest you read Fletcher's "The Myth of Jury Nullification" (IIRC, that's the title).
What needs to be done is the creation of a universal virtual architecture for the whole damn thing. I don't want to manage ten programs on twenty computers. I want there to be one program on the twenty computers that check in with another one to not only get work units / assignments but the cores too. So I can say to my controller box, computer 1 can run projects A,B,D,F,and Q, with weighted priorities of .1, .2, .4, .2, and .1 respectively. If no work is avalable, fall to next highest priority. Then have computer 2 work on A,C,E,and Q, randomly pick between A,C,E and only run Q if none of those projects have outstanding work (works for A,C,E being things like MOSIX).
The controller should also know the locations of the projects masters, so it acts as a data proxy unless otherwise allowed and can fetch newer versions of the cores in real time. Of course, the cores can just languish on disk until I sign them, should I chose to.
IOW, something like the now-abandoned distributed.net v3.
I'm wandering. Sorry.
_Knots
No, fat32 is not journaled. Beowulf_boy just thinks it is because Windows tends not to keep lots and lots of dirty data in RAM (as opposed to *nix, etc.) and so seems to have better crash-safety.
It's a lot like the mac's FS - HFS isn't journaled either, but since MacOS n (n 10.0) had single threading disk access and didn't keep a lot of dirty data in ram, the disk was usually in sync.
_Knots
So... if a blue laser (l = 440nm) has a better sampling resolution than a red one (l = 700nm), why couldn't a blue laser device read a disk burned by a red laser? Or would we have to get to UV lasers and use 350nm light?
TIA.
_Knots
TCP session hashes may be a clever solution to the problem as it now stands, but there's a so-called wheat-and-chaff secrecy system (don't have links, but google should be able to find something for the more curious) that is [almost?] completely undetectable to a listening party.
/dev/[u]random for its payload! Shift the key, and repeat. If the key runs out before transmission's completed, we can then agree on another key.
You have a document you want to send to me, but you worry that somebody's watching our TCP packets. Or worse - your document has the same MD5 hash as one of the multi-million songs out there by the RIAA (c'mon, there had to be an address-space collision sooner or later). We agree, using a key generation algorithm (ala SSH and others), to encrypt everything. Ok, cool - our packets now look like random noise addressed to each other. Now we make it an order of magnitude worse. We agree on a second key that is not used for encryption. We take this key, maybe do a transform on it, and then just use it as a flow-controlling shift register. If the current bit is one, we transmit a valid packet of data. If the current bit is zero, we transmit a valid packet that has part of
The power of this method is that, as long as the second key (well, really, shared-secret) generation is secure, the whole stream is nonsense. Any attempt to session-hash it is meaningless, and any except the hardest attempts to decode it will be difficult.
The method works better with smaller packets, obviously - smaller amount of data == more nondetermanism (from the point of view of one not in possession of the shared secret).
So maybe secrecy is our last and final option. And hey! If the hardware lockdown is really *that* anal, it would detect a PS/2 keyboard logger as a means of circumvention [can't have somebody capturing somebody else's passphrases]! =)
_Knots
RMS has his own motives, which you would better be asking him about. Ditto for everybody in the world, no?
As for my own motives, I see OSS as the best model to produce software - right in front of everybody, so the whole world can potentially learn, criticize (constructivly, ideally), or ignore. People certainly have the freedom to produce closed-source systems, BUT they do not have the right to do half of what they do with copyright these days (shrink-wrap EULAs) or force me to run proprietary software (gotta love open standards). That's just my view - proprietary software's cool; what I'm opposed to are abuses of copyright (preventing RE, the DMCA, the SSSCA, etc) as applied to software and generally crappy or sub-standard proprietary software (In my view, that includes Windows and virtually every piece of software to come from Microsoft, but not exclusive to MS).
If I see Windows running somewhere, I might laugh and tease about them using a joke OS, but hey, if it works for what they need it to, I'm not going to try to *make* them run Linux or BSD or MacOS or whatever - and people tease me for running Linux ('s fine), but they don't try to make me run Windows [corporations do, though, which is totally uncool].
Anyway, I'm wandering off topic.
_Knots
So what's so hard about this?!
If an arbitrary company doesn't want to use OSS, I'm not about to try to force them. Now, if it's the company I work at - remember, they pay me for my knowledge in this area, *I* am going to want to use Linux, but that's because I'm most comfortable with it. If they tell me I have to use Windows, I'm going to go down kicking and screaming. But a for-sale UNIX.... well, I probably could manage that, if they insisted.
Remember, OSS is a *gift* to the world - the world can choose to snub it if it so desires. There will always be somebody with the proverbial itch to scratch. I for one, will continue using Linux (and helping its development in my own small way).
Now, one area where this gets grey in my mind is companies like RedHat. They didn't develop the software, they just package it with pretty colors and sell support. So one could perhaps argue that they are just support contractors that *happen* to sell you physical media containing a free product already available online. (Hey, if that line of logic works, maybe we won't have to worry about companies violating the GPL - because suddenly they'd be responsible for *everything* in a code-tree that they sold.) But now what about cheapbytes and such? I suppose they're just selling physical media containing stuff already available on their FTP site [and really, they do just charge for media and shipping and little else, especially not the software] - so I guess that's not a problem.
_Knots
Oh no... between the DMCA and whoever owns the genes for human eyeballs, fingers, etc.... I'm going to be paying A TON to keep living. Though of course I'll be getting my Cease-and-Desist letters from the gene companies in jail because of the DMCA.
And I'll be stuck there until the media corps and the gene corps sue each other out of existence, fighting over my body parts' legality. Though they'll both agree that I can't have them.
_Knots
The DMCA effectively outlaws higher math. Think about it - data encryption is just complex higher math. Reversing it requires some skill with a data logger / oscilloscope, etc., and quite a bit of work, these days, shuttling mathematics around on paper. At least as far as I've seen.... maybe I'm wrong.
All somebody has to do to ban Calculus class forever from all highschools in the US is make some encryption based on integration or differentiation. I'm sure that's already happened.... I can see it now: "Calculus, the study of Differential Equations, and all of number theory have been declared unlawful because their primary purpose has become the circumvention of encryption."
[Moderators - it's supposed to be a joke. Mod appropriately]
_Knots