Because md5 is a *one-way* hash-function, a
server with a database of *md5 sums* for the
various chunks of aim.exe *could not in any way* be said to have a copy of aim.exe, or anything that could even be processed to produce aim.exe.
The sum is *one-way*.
The drawback here is that the space of offset x length pairs of md5 sums is much larger than the
size of the original program -- roughly the original program size squared. But if the aim.exe isn't *too* large, a single server with a coupla gigs of hard drive space might be able to do it.
Otherwise, the sum server could just use a cache of the recent values, and drop off the LRU.
This assumes that AOL gnereally doesn't change its
length-offset pairs that frequently.
(Oh - the chicken-and-egg problem: someone with a *legitamate* copy of aim.exe could seed the server with valid pairs. Remember, because md5 is *one-way*, doing so would not be comparable to exchanging copies of aim.exe.)
This case has a real bearing on the future of
that Assassination Politics guy, doesn't it?
I mean, isn't he in jail right now for *proposing*
something almost exactly like what the anti-abortion folks *did*?
Instead of verbal cheers, though, his proposal was
to actually send the assassins money to show his support. But again without any causal link...
Anyone out there thinking the same way I'm thinking? Isn't this relevant to his case?
Which is why end-to-end encryption is the way to
go. IPsec is the future... as wireless becomes
more pervasive, ubiquitous end-to-end encryption will become more and more obviously Right.
Dude, this is *unix*. You can convert *any*
file-at-a-time command-line utility to a batch
converter. How about:
for f in `find . -name "*.mp3"`; do
convert $f `basename $f.mp3`.ogg
done
(This is a wild hack, and it's for bash. Is Vorbis' extension really.ogg? I have no clue.)
If you're stuck on windows, you can still download the Cygwin tools and have a reasonable operating environment. Putting tools together is what unix is all about, man!
Actually, the data shows that the human genome
has been *infected* with bacteria DNA, which
has been incorporated into the genome -- I don't
believe that anyone knows whether these sequences
are active or not. I believe that the author
of this piece has his facts wrong -- but what are
facts?
This is completely tangential to the thrust of his argument, of course. There are a lot of odd genomic features which can yet be argues both pro
(look at all the junk! how could that be designed in?) and anti (general complexity: all our genes appear to interact in many more ways than previously thought) evolution. The jury's still out, and forever will be: remember the 'fake dinosaur skeletons' that the mice were building into the Earth Mark II in Hitchhiker's Guide to the Galaxy? There's no scientific argument that says that the same thing might not be done by an onmipotent God (once one has assumed the existence of such God -- but here our argument becomes circular).
In any case, the article author's "descended from bacteria" claim is a grossly inaccurate canard.
Actually, my girlfriend quite indignantly corrected me: "I got the joke! I know what Red Hat is!" My having owned (and in fact still owning) RHAT stock having something to do with this. =) I didn't even try to explain the
Alan Cox angle to her, though. =)
No, actually you probably want a 'stat' call.
To quote the man page: "access checks whether the process would be allowed to read, write or test for existence of the file (or other file system object) whose name is pathname." You actually
want to do the existence test, you don't want
to see if you can do it. Hence 'access' is
inappropriate.
'stat' may have more overhead than the 'open' call, depending on the implementation of your
filesystem. So it's probably just fine to
keep that 'open' call the way it is.
Ok, I just skipped Katz's review, for the obvious reasons.
Second review was pretty close on. I went in *wanting* to like the movie, and although the script was *definitely* clunky in the beginning, things starting improving and it did, indeed, become a fairly enjoyable PG-13 thriller. I *expect* to have to grit my teeth at inaccuracies in any movie remotely involving computers (even 'Pi'), and my teeth were surprisingly unworn out when leaving, and that made it all good. Also, I went with my non-techie girlfriend, so it was especially good to see the geeks win at the end. I think the movie captured the essentially idealism of free software pretty well. Even my girlfriend laughed and cheered in recognition when the heros start scrolling the stolen-and-supressed source code on the video broadcast. It was great. =)
Granted, I've also distributed DeCSS on flyers in Times Square in New York, so the idea appealed to me deeply. =)
Now the promised geek notes:
Yes, I did notice the red fedora on the spy cam and laughed out loud. I had to explain to my girlfriend that that was a very inside joke.
Was I the only one who groaned when the ex-cop said "dust the colon and backslash keys: those are the ones that geeks use and normal people don't"? My colon and backslash have been unused
(well, at the shell prompt at least)
since giving up DOS almost a decade ago. The shell prompt being used is definitely UNIXish,
and none of the commands typed on screen involve those evil characters. "*Forward-slash*!" I shouted, accidentally.
Ron Rivest's wife Gail was at the showing I attended. She noticed my unintended outburst above and had to make sure I knew that her husband was the 'R' in RSA (RSA Data Security actually heavily sponsored the film). I knew that, of course, and we had a pleasant discussion of the odd peculiarities of my PhD advisor, Martin Rinard, whom she knew well. =)
I thought that if I were a *true* geek, I'd be able to tell where the 10.X.X.X IP addresses used in the movie actually corresponded to. (Hint, not a satellite). I can't recall if they're a class-A private subnet or not -- I only use the class-B private subnets.
That's all the geekiness I can recall for now. =)
One other note, though: the movie skips quite a bit of plot at the very end --- actually, terribly typically for a PG-13 thriller. I'd love to hear people's ideas about *exactly* what happens after the Chinese dinner.
From the sponsoring company's web site, it looks
like they're targetting the telecommunications
industry which it appears still uses a significant
number of PDP-10s as telco switches. (Perhaps
"switch controllers" would be more appropriate?)
In any case, telcos tend to be *very conservative* about introducing new hw/sw, which is why they
may still be using their tried-and-true PDP-10
packages. Which is also why I don't see what the point of "new software" written in C would be:
if the telcos wanted to try new stuff, they'd be
trying new hardware, too. Wouldn't they?
But I don't *want* five different SyncML clients
on my memory-challenged Palm top. AvantGo backed
out of "OpenSource" MAL and left us with the two
slightly different MobileLink versions that
AvantGo and Vindigo use -- that's bad enough.
SyncML should not suffer this fate.
When will companies learn that it's the *server* that's valuable, not the piddly little SyncML client?
(FWIW: I'm perfectly capable of writing a Palm client for myself. My point is that *no one wants* five different slightly incompatible clients. We want one, officially supported, universally compatible client. And the SyncML folk aren't giving us that.)
Although there is a download entitled "WinPalmLinux" reference code, it doesn't
include anything remotely resembling
palmpilot client code. Does anyone know what
gives? Presumably the idea is that there will
be a "MobileLink"-type application to implement
the SyncML protocol on the client side? If so,
where is it?
Au contraire, Linux seems to be one of the primary
platforms supported by the SyncML reference code.
My guess would be that the SyncML folk are trying
to encourage servers to be written, and understand
that people *much* prefer to develop servers for Linux rather than windows.
This is a copy of I mail I just sent to rehmi@media.mit.edu.
do you really think you can build pengachu in volume for less than $50? It seems
to me that the memory resources alone (DRAM/Flash ROM) add up to more than
$50, and these are commodity parts whose price is unlikely to decrease
with anything other than time. Perhaps you should clarify that you mean
"$50 two years from now" which is very different from "$50 today".
Added to this objection is the electrical engineering rule of thumb I
learned as an undergraduate: a design's cost is roughly ten times its
component cost, once all manufacturing factors are added in.
As a data point, palm pilots based on a subset of your technology are sold
for $150 retail. But the devices sold for that price (which may well
translate into a $50 "direct to developing nation" cost) include only a
quarter of the memory pengachu does and none of the specialized media
hardware.
I'm curious for a cost justification.
--s
Electronic Election Requirements.
on
eLection '04
·
· Score: 2
There are three basic requirements for elections:
1. Inability for an observer to determine the vote of a particular voter.
2. Inability of the *voter* to prove his vote to an observer (special case of 1).
3. Ability of the voter to verify that his vote was included in the total.
4. Ability to prove that no non-voters were included in the total.
Current physical election systems give us 1 and 2, and to a lesser degree 4. (Compare voter rolls with totals). Electronic systems can (with strong cryptography) give us all four properties, *but* physical security of the voting place is still required to enforce 1 and 2 -- otherwise someone can look over the voter's shoulder. Failure of properties 1 and 2 opens the system up to vote buying and other fraud. Not a good thing.
So, I'd against "vote at home" scheme. But I would like to see electronic voting sceheme, because it would allow property 3. Every Florida voter would *love* to be able to verify that *their* vote was cast and counted, but in the current system this is impossible.
This is the letter I sent to my representative -- you may borrow from it, if you like, to help draft yours.
The chief of staff for Rep. Ernest Istook said his
proposed legislation to require mandatory internet
filtering in all federally funded schools and libraries
"seems like it's a no-brainer to the average Joe".
It may be, but the average Joe is wrong.
Internet filtering is not just ineffective -- see for example the
extensive reports written by the Censorware project at
http://www.censorware.org/ -- it is opposed to the most
fundamental principle of both schools and libraries: education
through free access to information. Surely we do not want to force
such critical assests as our libraries through *anyone's* sieve,
certainly not as undemocratic a sieve as most net filtering products
are (net filtering companies usually refuse to disclose even what
they are filtering to the public, much less submit their blinders to
democratic review). See the results of the censorware projects
"Foil the Filters Contest" at http://www.dfn.org/Alerts/contest.htm
for an idea of how pernicious this technology can be: every
filtering software product which Dick Armey recommends blocks
*his own web site* at the House of Representatives because it
contains the word "dick". Do we really want such blocks in every
school and library in the land?
Finally, such technological measures are destined to be ineffective
to those who really want to access pornography. The losers will be
the law-abiding citizen.
I hope that you will work to defeat this amendment
to bill HR4577 and work to convince your fellow legislators that,
despite the appearance of being a "no-brainer", this proposed
legislation is a danger to the very fabric of our democracy.
Sincerely,
C. Scott Ananian
MIT graduate student.
Use congress.org to locate and email your representative.
You both support the war on drugs, which
has swelled the American prison population with
hundreds of thousands of nonviolent offenders.
Both of you have faced questions about illegal
drug use, and the vice president has admitted
using illegal drugs. Yet neither of you has faced
prison time or arrest. First, how can you be a
credible force in the drug war considering your
history, and second what will you do to make
sure that the force of the law doesn't
disproportionately fall on the underprivileged?
Books such as Actual Innocence by Dwyer, Neufeld, and Scheck have documented the ways that provably innocent people have ended up on Death Row. If you believed innocent people had been executed, would you still support the death penalty?
Books such as Actual Innocence by Barry Siegel have
documented the ways that provably innocent people have ended up on Death Row. If you believed innocent people had been executed, would you still support the death penalty?
"free", huh? As someone who got arrested at the Republican National Convention in
Philadelphia for *exactly* "conspiracy to possibly
do something", I find that a very interesting
concept indeed.
Remember, the original cavitation missile article
stated that one of the big problems with these
guys are that you can't turn them: the missile's
shooting through a big vacuum (no air, no water),
so there's nothing for fins to push on. They
basically go in a straight line once they're
fired.
However, it may very well be that the missile
exploded before leaving the tube. That would
make sense.
This hoax really annoys me, because the board they describe *is* possible, and likely could be manufactured for even less than the price they made up. They giggle at "all the orders" they received, but this indicates a *real* business opportunity. It's not a joke. They offered a product people wanted and then laugh at people for wanting it -- fine. But to blather on about their made up technical specifications as if the whole project was ludicrously impossible to begin with -- well, that gets on my nerves.
Slashdot Mirror
The drawback here is that the space of offset x length pairs of md5 sums is much larger than the size of the original program -- roughly the original program size squared. But if the aim.exe isn't *too* large, a single server with a coupla gigs of hard drive space might be able to do it.
Otherwise, the sum server could just use a cache of the recent values, and drop off the LRU. This assumes that AOL gnereally doesn't change its length-offset pairs that frequently.
(Oh - the chicken-and-egg problem: someone with a *legitamate* copy of aim.exe could seed the server with valid pairs. Remember, because md5 is *one-way*, doing so would not be comparable to exchanging copies of aim.exe.)
This case has a real bearing on the future of that Assassination Politics guy, doesn't it? I mean, isn't he in jail right now for *proposing* something almost exactly like what the anti-abortion folks *did*?
Instead of verbal cheers, though, his proposal was to actually send the assassins money to show his support. But again without any causal link...
Anyone out there thinking the same way I'm thinking? Isn't this relevant to his case?
Which is why end-to-end encryption is the way to go. IPsec is the future... as wireless becomes more pervasive, ubiquitous end-to-end encryption will become more and more obviously Right.
for f in `find . -name "*.mp3"`; do .mp3`.ogg
convert $f `basename $f
done
(This is a wild hack, and it's for bash. Is Vorbis' extension really .ogg? I have no clue.)
If you're stuck on windows, you can still download the Cygwin tools and have a reasonable operating environment. Putting tools together is what unix is all about, man!
This is completely tangential to the thrust of his argument, of course. There are a lot of odd genomic features which can yet be argues both pro (look at all the junk! how could that be designed in?) and anti (general complexity: all our genes appear to interact in many more ways than previously thought) evolution. The jury's still out, and forever will be: remember the 'fake dinosaur skeletons' that the mice were building into the Earth Mark II in Hitchhiker's Guide to the Galaxy? There's no scientific argument that says that the same thing might not be done by an onmipotent God (once one has assumed the existence of such God -- but here our argument becomes circular).
In any case, the article author's "descended from bacteria" claim is a grossly inaccurate canard.
Actually, my girlfriend quite indignantly corrected me: "I got the joke! I know what Red Hat is!" My having owned (and in fact still owning) RHAT stock having something to do with this. =) I didn't even try to explain the Alan Cox angle to her, though. =)
'stat' may have more overhead than the 'open' call, depending on the implementation of your filesystem. So it's probably just fine to keep that 'open' call the way it is.
Second review was pretty close on. I went in *wanting* to like the movie, and although the script was *definitely* clunky in the beginning, things starting improving and it did, indeed, become a fairly enjoyable PG-13 thriller. I *expect* to have to grit my teeth at inaccuracies in any movie remotely involving computers (even 'Pi'), and my teeth were surprisingly unworn out when leaving, and that made it all good. Also, I went with my non-techie girlfriend, so it was especially good to see the geeks win at the end. I think the movie captured the essentially idealism of free software pretty well. Even my girlfriend laughed and cheered in recognition when the heros start scrolling the stolen-and-supressed source code on the video broadcast. It was great. =)
Granted, I've also distributed DeCSS on flyers in Times Square in New York, so the idea appealed to me deeply. =)
Now the promised geek notes:
That's all the geekiness I can recall for now. =) One other note, though: the movie skips quite a bit of plot at the very end --- actually, terribly typically for a PG-13 thriller. I'd love to hear people's ideas about *exactly* what happens after the Chinese dinner.
In any case, telcos tend to be *very conservative* about introducing new hw/sw, which is why they may still be using their tried-and-true PDP-10 packages. Which is also why I don't see what the point of "new software" written in C would be: if the telcos wanted to try new stuff, they'd be trying new hardware, too. Wouldn't they?
When will companies learn that it's the *server* that's valuable, not the piddly little SyncML client?
(FWIW: I'm perfectly capable of writing a Palm client for myself. My point is that *no one wants* five different slightly incompatible clients. We want one, officially supported, universally compatible client. And the SyncML folk aren't giving us that.)
Although there is a download entitled "WinPalmLinux" reference code, it doesn't include anything remotely resembling palmpilot client code. Does anyone know what gives? Presumably the idea is that there will be a "MobileLink"-type application to implement the SyncML protocol on the client side? If so, where is it?
My guess would be that the SyncML folk are trying to encourage servers to be written, and understand that people *much* prefer to develop servers for Linux rather than windows.
But how does this affect Apple's "supercomputer" advertising taglines?
There are three basic requirements for elections:
1. Inability for an observer to determine the vote of a particular voter.
2. Inability of the *voter* to prove his vote to an observer (special case of 1).
3. Ability of the voter to verify that his vote was included in the total.
4. Ability to prove that no non-voters were included in the total.
Current physical election systems give us 1 and 2, and to a lesser degree 4. (Compare voter rolls with totals). Electronic systems can (with strong cryptography) give us all four properties, *but* physical security of the voting place is still required to enforce 1 and 2 -- otherwise someone can look over the voter's shoulder. Failure of properties 1 and 2 opens the system up to vote buying and other fraud. Not a good thing.
So, I'd against "vote at home" scheme. But I would like to see electronic voting sceheme, because it would allow property 3. Every Florida voter would *love* to be able to verify that *their* vote was cast and counted, but in the current system this is impossible.
You both support the war on drugs, which has swelled the American prison population with hundreds of thousands of nonviolent offenders. Both of you have faced questions about illegal drug use, and the vice president has admitted using illegal drugs. Yet neither of you has faced prison time or arrest. First, how can you be a credible force in the drug war considering your history, and second what will you do to make sure that the force of the law doesn't disproportionately fall on the underprivileged?
Books such as Actual Innocence by Dwyer, Neufeld, and Scheck have documented the ways that provably innocent people have ended up on Death Row. If you believed innocent people had been executed, would you still support the death penalty?
Oops, my bad; I got the author of the book (and therefore the link) wrong. The following is the corrected question. Sorry. =(
Books such as Actual Innocence by Barry Siegel have documented the ways that provably innocent people have ended up on Death Row. If you believed innocent people had been executed, would you still support the death penalty?
"free", huh? As someone who got arrested at the Republican National Convention in Philadelphia for *exactly* "conspiracy to possibly do something", I find that a very interesting concept indeed.
However, it may very well be that the missile exploded before leaving the tube. That would make sense.
There are laws against false advertising.
Wrote the webmaster.
This is a simple ISP fuckup. Telecom New Zealand screwed up.
And here's the start of the apologies. Paul Vixie apologizes, even. They all shake hands. Well, maybe not really, but still:
The story as reported is all lies and misinformation.