(please bear with my probably ignorance of trade secret case law and feel free to jump in and correct me)
it just dawned on me - what we're really talking about here is the key that was found in the Xing code, plus some others that were guessed once it was known.
It's pretty obvious that the guessed ones can't be copyrighted.... being encrypted they haven't even been published so they must be trade secrets.... the only people who could be prosecuted would be people who gave them away... and no one did, as I understand it, they were obtained by deduction, not reverse engineering or any sort of theft.
On the other hand the original key that was in plaintext in the Xing code was obtained from a copyrighted work..... can something that's just a number be copyrighted? (I know it can't be trademarked)? Can I copyright '9', or better yet '0'? how about '42'? or '78687622'? can one draw the line (yeah I know it's law, not math and they have different criteria for getting stuff done.... but if some people are going to own numbers I want to stake out some usefull ones before the land rush gets going).
Seriously though it seems that one could 'own' a number in a certain context.... for example '987492837498234898732' in the context of DVD might be a copyrightable number.... but what if I want to use it in a paper on number theory? or it's a large prime I want to use in a paper on primes?
Of course the problem here is that in order for this DVD-number to be usefull it has to be secret but copyright requires publishing - the minute it's published the number loses all it's worth - so maybe we're just looking at lawyer smoke&mirrors
is always a bad idea.... it just makes the other users of the stable angry.
Leaving the key under the mat is also a bad idea
Letting amateurs implement crypto doesn't work so well either
But the worst mistake is to alienate a whole bunch of smart people who understand locks by selling them horses but not letting them ride them
In the future if you are basing a business on the use of large secret numbers you had better use really big ones.... and maybe not leave them around where people can find them
well by 'technology' I really mean 'fab technology' and 'design technology'..... how do you build lots of them cheaply? how do you design them? (CAD tools are expensive and existing designers have a lot of money and mindshare invested in what they are using today) etc etc
Being able to build a few gates in the lab is one thing.... getting it to a mass market and cheaper than the existing technologies with all the momentum they have going for them is really hard. Not impossible mind you..... just harder than most people think
you add encryption in as an afterthought.... way back when DVD was about to come out they decided to toss in encryption quite late in the game.... oops.... it was done quickly and I'd guess (mostly) by people who were video compression weanies rather than crypto-weanies...
Of course we all know that the only good crypto system is one that's been through lots of public exposure and attacks.... and passed.... all that's happened really is that the DVD one's got it's public exposure and failed.... they just shouldn't have released it before it was proven.
And remember this isn't the first DVD crack.... others have involved subverting Win95 to get at unencrypted bits on their way to the screen
yup - a molecule does not a technology make - any more than 'silicon' without the manufacturing processes to etch it, dope it, drop metal and insolators on it is just refined sand
What is needed is an infrastructure that goes on top of this.... how do you put the molecules into the states you want, how do you sense their state, how do you get that information to the outside world.
And once you get it out what do you do with it? put it into a molecular computer?.... then probably you're using the same technologies you fabbed the ram with.... put it into a 'traditional' silicon computer? then you need a whole other bunch of technologies/infrastructure that allows these molecular structures to be fabbed alongside silicon ones....
In other words there's a lot of work to be done!
It'll happen one day.... probably not next week, or next year
Oh yeah and noise/cosmic rays/quantum effects etc etc you have to be able to handle all those other things that can cause these molecules to change state when you don't want them to - with ram you can do heavy ECC and scrubbing to get reliability.... for random logic in, for example, a CPU it's a whole different, and much harder, problem
I think that it's main competitors (VCS or cadence) go for lists of $40k/license. A lot more people use these tools rather than modelsim. There's also a bunch of cheapo VFerilog implementations that I don't know of anyone using for real work.
This is one area that the open source model isn't going to play quite so easily - basicly because there's 0 tolerance for bugs when you're taping out a $500k chip - people are very conservative about what they use - look how long it took for VCS to catch on (and it was probably 20 times faster than the interpreted standard of its day).
But remember that these numbers are also 'list' price - there's a LOT of discounting goes on - hit the sales guy up at the end of the quarter when he's up against his quota, or bundle it with some other purchase and you can get a much better price. Oh yeah and make SURE they see that you're trying out the competitor's product:-)
yup - I'm a daily verilog user (and a one time Verilog compiler writer many years back.... anyone want a compiler for their 68k mac:-) - we buy the biggest, meanest boxes to run our stuff on - they're cheap compared with the software. However some things about Verilog performance:
little locality of reference - big TLBs are important - caches of any size don't seem to be small enough (except for reall small simulations - think of it this way - every clock tick everything in memory has to wake up and do something)
as a result how fast your main memory subsystem is is very important
Beowulf is out - Verilog is amenable to a couple of solution spaces which beowulf is sort of in the middle of:
tight SMP (think of Verilog as lots of tiny processes)
farms
most modern piped architectures suck because there's lots of pointer chasing underlying the simulation engine
it probably doesn't matter much how many registers your CPU has - most of the time is not spent calculating expressions but in doing the business of simulation
actually using the LSBs of music CD data as a one-time pad is an interesting possibility.... after all the pad distribution problem is solved for you:-)
The down side of course is that while there are an awfull lot of CDs in circulation.... therevreally aren't that many.... I can imagine if it got to be common there'd be some poor shmo at the NSA who spent all day trying to break into the impossible celophane that imprisons CDs ad tossing them in a CD drive to be read
Of course eventually the work of some bands would be prized for how close their music comes to random noise.... then finally we'd just give up on any pretense of music.... "excuse me could you please direct me to the white noise bins thankyou"
nah - one usefull molecule does not a technology make - there's lots of interesting things people have proposed that you can do at the molecular level. Much more interesting are things like:
how do you fab them? in bulk, with what defect densities?
what are their operating temp ranges?
how do you hook them together to do usefull stuff?
how do they talk to the outside world? (do they have to be fabbed in a Si world in order to get hooked up to usefull things like PCI buses etc)
>.WHEN CAN WE USE BEOWULF ON IT!?!?! as things get smaller, and clock speeds go up, and the speed of light doesn't change.... non-local communication becomes more and the limiting factor in how complex an object you can design. One way to get some life out of your designs is to move into the 3rd dimension (start layering devices on top of each other - something that is probably much more doable using a nanotech-style assembler than the current mask/etch assembly methods.... sadly you only get to do this once..... given that there are not a lot of available, easily accesable spare dimensions lying around).
The first thing that goes is a global clock.
It gets replaced by a bunch of local clocks - or in essence a bunch of discrete CPUs communicating with their 2d (or 3D) neighbors. Or you toss the clocks altogether and do only asynchronous logic (something that's still really in the academic curiosity realm - but promises faster, cooler logic - and chips that automatically clock-chip themselves when they get colder).
To handle really big problems you're going to have to write the software to spread the problem over the multiple clock domains (ala Beowulf) so you might not have a choice in the matter
every other year someone shows up with the wonderfull new technology that's gonna eat silicon's lunch..... but there's just so much invested in Si R&D and so much practical process knowledge about how to build Si structures that it has just so much momentum and nothing else has managed to catch up with it. Plus there's a whole bunch invested in CAd tools, and training chip designers like me
One day there will be something, and it will almost certainly be at the nanotech level where all the existing stuff breaks down.... but it's going to be a long ugly transition.... Si will hold on as long as it possibly can (kind of like how modem tech kept coming back after people predicted it' demise) I bet longer than most people predict.... and molecular level stuff's going to be initially unreliable (and probably get a bad rep as a result).
Me I'm still hoping for the non-electrical nanotech stuff and assemblers so we can get away from this 2d chip paradigm we're currently stuck in
You'd think that by now the people running these marketting campaigns would realise that 5-10% of their target market (let's face most of the people they are selling to are male so it's closer to 10%) is red-green color-blind
Mmmm..... Pentium-muddy-brown...... it must be fast....
It's not just my name and address - that was a particularly annoying example.
If I visit Ford's web site and Ford banner ads follow me all over the web for the next week or two (purely an example, unlike the Dell thing, this hasn't actually happened to me) - then again someone's trafficing in my identity and information about my web usage - they may not know exactly who I am - but it doesn't matter they are still passing information around about me, and money is probably changing hands - and I WANT MY CUT!:-)
Of course they can figure it out - lots of ways - for example - they can take my IP address, do a reverse lookup on it, discover my domain name - look at the whois record from the internic - and send mail to the administrator (me - I get regular paper spam from the various domain names I administer).
Much more likely though, you're right, the correlated something from my visit - maybe a cookie someone else left in my browser, or my IP address - with some other place where I did reveal my identity - chances are they sold it for money to Dell - and as I said I WANT MY CUT
Slashdot Mirror
it just dawned on me - what we're really talking about here is the key that was found in the Xing code, plus some others that were guessed once it was known.
It's pretty obvious that the guessed ones can't be copyrighted .... being encrypted they haven't even been published so they must be trade secrets .... the only people who could be prosecuted would be people who gave them away ... and no one did, as I understand it, they were obtained by deduction, not reverse engineering or any sort of theft.
On the other hand the original key that was in plaintext in the Xing code was obtained from a copyrighted work ..... can something that's just a number be copyrighted? (I know it can't be trademarked)? Can I copyright '9', or better yet '0'? how about '42'? or '78687622'? can one draw the line (yeah I know it's law, not math and they have different criteria for getting stuff done .... but if some people are going to own numbers I want to stake out some usefull ones before the land rush gets going).
Seriously though it seems that one could 'own' a number in a certain context .... for example '987492837498234898732' in the context of DVD might be a copyrightable number .... but what if I want to use it in a paper on number theory? or it's a large prime I want to use in a paper on primes?
Of course the problem here is that in order for this DVD-number to be usefull it has to be secret but copyright requires publishing - the minute it's published the number loses all it's worth - so maybe we're just looking at lawyer smoke&mirrors
Leaving the key under the mat is also a bad idea
Letting amateurs implement crypto doesn't work so well either
But the worst mistake is to alienate a whole bunch of smart people who understand locks by selling them horses but not letting them ride them
In the future if you are basing a business on the use of large secret numbers you had better use really big ones .... and maybe not leave them around where people can find them
Being able to build a few gates in the lab is one thing .... getting it to a mass market and cheaper than the existing technologies with all the momentum they have going for them is really hard. Not impossible mind you ..... just harder than most people think
Of course we all know that the only good crypto system is one that's been through lots of public exposure and attacks .... and passed .... all that's happened really is that the DVD one's got it's public exposure and failed .... they just shouldn't have released it before it was proven.
And remember this isn't the first DVD crack .... others have involved subverting Win95 to get at unencrypted bits on their way to the screen
al connections between swimming in milk and driving an SUV in traffic
What is needed is an infrastructure that goes on top of this .... how do you put the molecules into the states you want, how do you sense their state, how do you get that information to the outside world.
And once you get it out what do you do with it? put it into a molecular computer? .... then probably you're using the same technologies you fabbed the ram with .... put it into a 'traditional' silicon computer? then you need a whole other bunch of technologies/infrastructure that allows these molecular structures to be fabbed alongside silicon ones ....
In other words there's a lot of work to be done!
It'll happen one day .... probably not next week, or next year
Oh yeah and noise/cosmic rays/quantum effects etc etc you have to be able to handle all those other things that can cause these molecules to change state when you don't want them to - with ram you can do heavy ECC and scrubbing to get reliability .... for random logic in, for example, a CPU it's a whole different, and much harder, problem
> hit is just overwhelming.
well maybe either is cool but the combination is kind of pointless ..... after the bomb (and the EMP) - who are you going to talk to ....
This is one area that the open source model isn't going to play quite so easily - basicly because there's 0 tolerance for bugs when you're taping out a $500k chip - people are very conservative about what they use - look how long it took for VCS to catch on (and it was probably 20 times faster than the interpreted standard of its day).
But remember that these numbers are also 'list' price - there's a LOT of discounting goes on - hit the sales guy up at the end of the quarter when he's up against his quota, or bundle it with some other purchase and you can get a much better price. Oh yeah and make SURE they see that you're trying out the competitor's product :-)
at $20k/license this is one thing you go out and buy that cyro-athlon to run on ....
The down side of course is that while there are an awfull lot of CDs in circulation .... therevreally aren't that many .... I can imagine if it got to be common there'd be some poor shmo at the NSA who spent all day trying to break into the impossible celophane that imprisons CDs ad tossing them in a CD drive to be read
Of course eventually the work of some bands would be prized for how close their music comes to random noise .... then finally we'd just give up on any pretense of music .... "excuse me could you please direct me to the white noise bins thankyou"
Yeh that's what I was refering to when I mentioned 'non-electrical systems' ..... rod-logic computers and that sort of stuff .....
The first thing that goes is a global clock.
It gets replaced by a bunch of local clocks - or in essence a bunch of discrete CPUs communicating with their 2d (or 3D) neighbors. Or you toss the clocks altogether and do only asynchronous logic (something that's still really in the academic curiosity realm - but promises faster, cooler logic - and chips that automatically clock-chip themselves when they get colder).
To handle really big problems you're going to have to write the software to spread the problem over the multiple clock domains (ala Beowulf) so you might not have a choice in the matter
One day there will be something, and it will almost certainly be at the nanotech level where all the existing stuff breaks down .... but it's going to be a long ugly transition .... Si will hold on as long as it possibly can (kind of like how modem tech kept coming back after people predicted it' demise) I bet longer than most people predict .... and molecular level stuff's going to be initially unreliable (and probably get a bad rep as a result).
Me I'm still hoping for the non-electrical nanotech stuff and assemblers so we can get away from this 2d chip paradigm we're currently stuck in
for the Mountain Dew/Jolt vending machine with the biometric sensors .... customer approaching ..... notes: glazed eyes, overweight, pizza stains, slurred speech, carpal tunnel ........ deduces: hacker and/or gamer ..... result: ca-ching increase price $1 ..... wait .....notes: boxter keychain ..... deduces: recent IPO .... result: ca-ching, get it while it's hot before the stock goes down, increase price $5
oh .... but you have to include a URL to this slashddot thread ....... just so they can see what PR-dummies they are ....
ummmm ..... crap in the corner?
Don't forget your TiVo box is running linux .... you should have it cracking RC5 between fields ..... (and while you're sleeping ....)
but when the aliens use their super-sekrit farfenugen sucking ray - you're screwed ....
Mmmm ..... Pentium-muddy-brown ...... it must be fast ....
don't hold your breath - the memory runs at the same speed - you're going to need DRAMS that reliably go that fast too
that's scary .... I have 4 of these .... :-)
If I visit Ford's web site and Ford banner ads follow me all over the web for the next week or two (purely an example, unlike the Dell thing, this hasn't actually happened to me) - then again someone's trafficing in my identity and information about my web usage - they may not know exactly who I am - but it doesn't matter they are still passing information around about me, and money is probably changing hands - and I WANT MY CUT! :-)
Much more likely though, you're right, the correlated something from my visit - maybe a cookie someone else left in my browser, or my IP address - with some other place where I did reveal my identity - chances are they sold it for money to Dell - and as I said I WANT MY CUT