I also have a question for those in the US. Can you buy single P4 chipsets without Rdram?
Yes, and it would be an ilegal tied sale if Intel tried to require it.
Intel seem to have screwed up really badly here. RDRAM is going to remain expensive, not least because RAMBUS is suing all the fabs with their smurfed SDRAM 'patent'.
A billion dollars of RAMBUS stock may sound a lot but that is nothing to losing 2% of market share to AMD. It appears to me that Intel have stalled the P4 range entirely with this deal.
Plus the RAMBUS stock that Intel will get is probably worth much less than a billion dollars right now, after the SDRAM fraud finding RAMBUS has been in free fall - remember that the judge only reversed some of the fraud findings and did not reverse the non infringement ruling.
The genius that made this mess probably gets paid several million a year with millions of options on top. After this fiasco he probably wishes they were puts not calls.
The folks yammering for W3C to come in and support Netscape's API should know that the reason Microsoft and Netscape are both obsoleting the interface is that the W3C DOM and Object Linking and Embedding specs replace them.
Not that Netscape ever played the standards game honestly. Their idea of a standards submission was to fax a copy of the spec to MIT a few hours before they released the code (yes it was a fax, not even an email!).
Netscape spent a lot of PR bucks claiming that Marc Andressen was the inventor of the Web, they even had a PR flack write a book 'Archotects of the Web' that gave Tim Berners-Lee (the true inventor) less than four paragraphs. W3C and MIT set up Microsoft to screw Netscape in return.
By the way Sun did not 'win' the Java suit, they accepted Microsoft's offer of a settlement. Clearly Microsoft understood the consequences of the settlement better than Sun. Like be carefull what you ask for, you might get it.
Actually you can. Ever wondered why Neutrasweet is still under patent after all these years? Their lawyers bought enough congressmen to get a bill passed to extend the life of their patent.
The drug companies bought a similar extension while the GOP had the Senate. They might not have sold anything as wicked as pardons to tax dodgers but they certainly sold legislation.
I would consider it hilarious too that you are preferring a less secure operating system than the one you used to ridicule for it security
Unless you are talking about one of the striped UNIX distributions with all the non-essentials taken out neither UNIX nor Windows NT is secure.
UNIX was certainly not designed to be secure as many would have people believe. Security consists of a bunch of ad-hoc add ons. Windows NT does have a security architecture in the kernel, pity that the authors of IIS had such a fondness for toxic feature sets like active code.
/. is no longer the pure nest of "rabid Linux-ism" that it used to be.
That is true, I got to karma cap in 50 posts and I am a veteran member of the MIT AI lab UNIX-Haters list. And that ran long before Windows NT 3.5 came out.
Actually it is hilarious to see items from the Unix haters list getting recycled as anti-Microsoft items. Security in particular, Richie and co barely had a clue and as for Eric Allman (author of Sendmail) his software still holds the all time record for CERT advisories of any application.
There is actually a significant difference between active-x and plug ins. The default security settings for the browsers are to require active-x code to be signed. Plug-ins are just another executable to download off the net and install, most are unsigned and because there is no built in download proceedure there is nothing to guide the user to require them to be signed.
Both active-x and a Netscape plug-in have unlimited ability to screw up your machine to the full extent of your user privs - which on a Windows 98 box is root. A signed active-x control does at least come with an assurance that the signer has been authenticated and has undertaken not to include malicious code and that the control has not been modified since. That is not perfect security, but it is a lot better than you get without the signature. If the signer does screw you up you have evidence that proves what he did and a contractual undertaking from him to wave about in court.
If this company is truly steamrolling gates & co. in court, they probably have a valid patent.
They are not. Cringely filters his entire world view through the lens of Microsoft hatred. It is like asking one of the people throwing bombs or shooting protesters in the middle east for an accurate assesment of the 'peace' process.
The Eolas patent is utterly bogus. There is racks of prior art for using embedded applets in the Web. There were embedded applets before NCSA even began work on Mosaic.
Since the Eolas case has been in the discovery stage for years it is difficult to see how an outside observer could make any accurate statement on the proceedings. The only rullings that would be heard to date would be the standard motions to dismiss that are always filled.
Fact is that Eolas probably don't want to allow the case to go to court because if they lose they lose big, their patent is busted to smithereens.
Another fact is that with the Republicans keeping Clinton's 'liberal' judges off the bench and the Democrats giving the GOP a taste of its own medicine in return the number of patent cases heard each year is actually declining slightly even though the number filled is exploding. At the rate things are going patent law will be moot because by the time a case is heard the patent will have expired. Hooray!
Actually you guys are both off base. Apparently Jobs didn't put a clause in the agreement to stop Gates shorting their stock - which is precisely what they then did.
So Microsoft owned a lot of apple but they also owed rather a lot of apple shares. Smart move.
It would be interesting to know if the guy has contacted Microsoft. Many of the features of D are in C# and vice versa. It would be a good thing if the future of programming languages was not considered the domain of Microsoft alone (bad) or Suns lawyers alone (worse).
Many of the features look pretty sensisble. There is now pretty unanimous support for dropping Multiple inheritance. The problem with multiple inheritance being that it leads to programs only the original authors understand.
It is disapointing that the syntax was not changed more radically. I for one am pretty bored with typing semicolons at the end of lines. Using braces for block structure is equally tedious.
The garbage collector is of the 'stop everything and collect' type, this is not a good scheme as anyone who has seen a PET running Microsoft Basic GC will agree. The incremental GC in.NET is a better scheme, even if it is slower overall. But that is an implementation detail.
It would be good if people would start to look at adding support for parallel program execution. The threads programming model is very clunky and hard to use, in part because there is no means for the program to perform checks on access conflicts.
Also a persistence model should be part of any new language. The current division between programming language and database is a lot of wasted overhead.
DES has been cracked by brute force. Never cracked via cryptanalytic means. I suggest you start paying attention to detail.
I suggest that you take your own advice and pay attention to detail yourself.
I said that 3DES is simply an extension of a broken cipher. Brute force is a perfectly respectable attack.
Or was your attempt to frame the argument that way simple dishonesty?
There have in fact been several attacks against DES that have lower complexity than brute force, however in practice the trivial parallelism and lower complexity of brute force tends to win. The fact that nobody has built a machine to implement Adi's attacks is irrelevant. The fact that the AES contenders were designed with the knowledge of Adi's recent techniques and DES was not is significant.
The key size of DES was reduced to 56 bits for a good reason, to ensure that the aparent strength of the cipher matched the actual strength. That may not be a big thing to you, in the cryptography community it is.
It is pretty easy to 'win' an argument like this on slashdot where most of the posters are like yourself journeymen at best and do not have the internal knowledge of the field. However you are going to find it much harder in the group you aspire to call your peers.
Oh, and if you think this is flaming, I suggest you get on the wrong end of an argument with Phil Z. or Bruce S.
Because Sun simply doesn't have the exclusive power over Java that Microsoft has over any of their technologies
Odd thing is that Microsoft is explicitly giving up control over C# in its current proposal to the ECMA wheras Sun insisted that it have sole control over the future direction of Java.
If the Open Source Community wants to have an open standards based successor to C then C# looks a better bet than Java at this point.
dotNet is not going to kill Apache for the simple reason that the concept of dotNet requires the ability to take existing applications and wrap a SOAP front end arround them so that other dotNET applications can access their resources.
In the very near future Open Source is going to become a serious threat to Sun's survival. Already Linux is taking serious market share from Sun for the same reason that Office Max furniture is taking market share from Herman Miller's Aeron chair. VC capital is harder to come by and much more impressed by a rack of cheap Linux boxes than a high end Solaris server.
For Sun to survive they have to make sure that their platform is too complex and rapidly changing for the Open Source movement to keep up while all the time pretending that they are on their side.
You are wrong, DES has been broken repeatedly. The world record for breaking DES is held by John Gilmor and Paul Kocher who built custom hardware to break it in less than two days. You can buy the book from Amazon. Or you can read the EFF press release
People might take you a bit more seriously as a 'Senior Security Specialist' if you bothered to read some of the basic litterature in the field before advising people with rather more knowledge, experience and standing in the field than yourself to "have your doctor up your antipsychotic medication".
3DES has not been broken to date. However it is 2^16 times less secure than other ciphers that have been equally intensively studied, execute considerably faster and require less keying material. That makes it a bad cipher according to most of the generally accepted definitions in the field - which by the way you will find in Bruce's book as well.
Anyone who compiles plaintext databases of credit card numbers on machines connected to the Internet is asking for trouble.
Unfortunately this will never be possible unless the USPTO ends its current corrupt scheme.
At present there is no means by which anyone can know what patent applications are in progress. It appears that even the modest changes under which patent applications would be published a year after filling are now 'optional'. What this means is that a corrupt applicant can easily file for a patent on somebody else's invention.
Another common trick is to extend an existing patent to include a claim on somebody else's invention that was published after the original application. Under the current rules the applicant gets the benefit of the original filling date.
Patent fraud is real and its legal. It is even encouraged by the USPTO.
They should let the competitors take anything they like. The only drug that would be worth a competitor taking would have to improve mental concentration over long periods of time without causing serious mental problems in the long run.
So far nobody has found anything that does not have profound negative side effects.
A chess master might get an advantage from taking Ritalin in one constest but even that is doubtful. It is a very powerful drug that messes with several areas of the brain. If you are a chess grandmaster you almost certainly don't suffer from ADD. Like if you had ADD how did you get there? You might well suffer from dyslexia which is the problem that many Ritalin victims actually have it being easier to prescribe an addictive drug than diagnose the problem.
One of the ironic things about high school and college abuse of Ritalin is that the kids who take it so that they can concentrate hours at a stretch would almost certainly learn more if they took regular breaks. Fatigue is the brain's signal to the body that it needs a rest. Transfer from short term to long term memory appears to work best as a background task while the variables are not locked by another process. It is better to take regular five minute breaks than try to sit down for hours at a stretch.
LSD may have helped the Beatles write Sgt Pepper but I don't think anyone considers that it would have helped their creativity over the long term if they had kept taking it.
On the contrary. I'm in the field, and I regard him as part of the very front rank. I wouldn't say he's another Coppersmith
In other word you would not put Bruce on the same level as Coppersmith, Shamir, Rivest, Rogaway and so on, or as I put it not of the very front rank. Bruce is the Issac Azimov of cryptography, not its Einstein or Newton.
It is somewhat rich for Bruce to imply in 'Secrets and Lies' that he has suddenly discovered that security is about risk control not risk elimination. If he has only just realised that then he should probably give me credit for putting him straight since I pointed out precisely that point to him when we talked at RSA some years back. Not that I was the first to think of it by a long way.
Slow, yes. Susceptible to a meet-in-the-middle, no. Schneier, 12.3: "[If DES were a group], DES would be vulnerable to a meet-in-the-middle known-plaintext attack that runs in only 2**28 steps".
DES is not vulnerable to a meet in the middle attack but 3DES is in such a way that the complexity of breaking 3DES is only twice that of breaking DES, despite having three times the key length. That is what makes it a bad cipher, the fastest known attack is well short of brute forcing the keyspace.
The details of the attack are discussed in AP with respect to 2DES, to break 2DES you simply construct an in-memory table of encrypting forwards from the known plaintext (cost = O(2^56)), construct another backward from the known ciphertext (cost = O(2^56)) and look for a match (cost = O(2^56)), total cost = O(2^56).
The attack can be extended to 3DES at the cost of performing two steps together, giving overall compexity O(2^112). It is a very well known result in the field and one the reason why those in the know are depricating 3DES, it is not a good cipher, it is merely an extension of a previously broken cipher.
Please point me in the direction of an implementable protocol which does provide perfect repudiability.
None gives perfect non-repudation, however PGP is designed to give pretty good PRIVACY even when the participants are pseudo-anonmous. It does not attempt to support a legal infrastructure, allow parties to place legally enforceable constraints on the liabilities they incur in authenticating a keyholder. As a result PGP is widely used amongst geeks but has a very limmited enterprise use. The vast majority of RFPs issued stipulate a PKIX conformant PKI.
I know Phil. He's one of the lowest-maintenance people I've ever met. Friendly as all get out, and patient with newbies. Would you care to enlighten me as to his ``combustible'' nature?
He has mellowed considerably since the FBI got off his case. However when the PEM vs. PGP war broke out, which is the time in question Phil was definitely of combustible nature. The FBI certainly did not help, but were certainly not the original cause.
Unfortunately rather than simply fix the parts of PEM that were monumentaly broken (the hierarchical CA system) Phil introduced competing formats all the way along the line.
There are 100 million email clients that ship with high quality crypto built in. However rather than leverage that deployed base you and the rest of the OpenPGP community spend your time explaining to people why they shouldn't use it.
Bruce is one of the best known experts, however he is not regarded in the field as being of the very front rank. One of the reasons for this being his habit of issuing 'Schneier-grams' which tend to make half-assed critiques of other peoples work in scathing tones only to be dropped quietly sometime later when his argument is shot down. His IPSEC critique was not exactly his best move.
It is always easier to state algorithms to steer clear of than ones to rely on. At this point IDEA is somewhat suspect, but when Applied Crypto 1 came out it was actually the best 128 bit cipher then available.
At this point most people are recommending AES (nee RIJNDAEL). The only reason to use 3DES is if you are forced to, there are still many banking applications that mandate DES. But 3DES is not a good cipher, it is slow and is subject to a meet in the middle attack that means that you do 3 times 56 bits of work to get 112 bits of security.
As far as software goes, practically all mail agents have S/MIME support built in. As far as security goes there are no serious attacks known against either S/MIME or PGP, beyond the fact that the chuckleheads in both IETF working groups flubbed the encryption of the subject line in both cases.
One problem with PGP is that it only really works well for confidentiality. It does not handle non-repudiation too well. Alice may know the message comes from Bob but proving it in court would be rather hard. Trusted Third Parties do have their uses.
The other technical problem with PGP is that it depends on the users being technically competent which most people are not.
The non-technical problem with PGP is the somewhat combustible nature of Phil Zimmerman. He is somewhat high maintenance. There is no reason why S/MIME and PGP use entirely different packaging formats except Phil's NIH policy, somewhat sad. The result being that Microsoft, Netscape, Lotus etc. implemented S/MIME and not PGP.
More recently the stale PKI/PGP debate has been rendered obsolete by technology such as XKMS which allows a client to use any PKI back end at all and not have to worry about how it works or how to configure it.
There are several good reasons not to. First the SSC was abandoned and the tunnel filled in. The cost of building it would be back to the original price.
Second, the SSC would have cost a lot more than $2billion to complete in the first place. Try closer to $8billion, plus the running costs.
Third CERN have already started building the LHC which does the same science at a fraction of the cost and will certainly be completed first. The SSC had nothing to do with science, it was putting the US flag on the thing that was the whole point of the exercise
Fourth, the site chosen was a dump, a redneck dry county where the most intelligent natives are the numerous fireants.
My point was that WEP does not support rekeying. It should support rekeying, it should support many things.
The second set of papers have demolished the proposed fixes that nobody has implemented. I doubt that a workarround will be necessary since that set of proposals is now completely dead.
It seems to me that low volume wireless LANs are pretty safe, and can be completely safe if they rekey on a regular basis.
Read the paper. It does not matter how often you rekey or whether you buy the 40bit or 128 bit cards. The algorithm used is a stream cipher and will XOR your plaintext with one of 2^24 ciphertext streams that are generated from your key.
The attacker can cause the gateway to act as an oracle for any given ciphertext stream.
If you rekeyed every hour you would be safe (ish). However the WEP protocol does not support rekeying and everyone in the network has to use the same key. So you would have to update all your machines manually constantly.
Towards the end of Picasso's life he started to find that people stopped cashing his checks. The owner of the local grocery store had discovered that he could sell a Picasso check which of course had Picasso's signature on it for considerably more than the cost of the groceries.
Don ain't exactly Picasso, but I'll bet that his signature would fetch more than $2.56 on ebay, paticularly if it was on one of his famous hexadecimal checks...
No, the Berkeley attack broke WEP version 1, the new attack breaks the proposed fix as well.
The big problem with the 802.11b folk is that in the beginning they had no security people and now they only have a couple and won't actually let them do what needs to be done.
The original WEP protocol was secure as reviewed by the NSA, then they substituted a stream cipher for the block cipher for better performance, completely breaking the scheme. Truncated IVs are not a serious problem with DES, plenty of protocols use them. Truncating the IV utterly destroys the security of RC4.
The deeper problem is that WEP attempts to provide 'equivalent privacy' to ethernet. But a wired network does not just provide some privacy it provides authentication. The big problem with WEP 1 or 2 is that there is no way to stop a fired employee surfing from the car park.
At present the (sensible) companies that are deploying 802.11b on a large scale are wrapping IPSEC arround it.
The best way to solve the problem however is to fix the protocol itself, and use a different key for each card instead of the same key for every card in the network. The 802.11b chumps keep rejecting this idea because it prevents the use of broadcast - the idea of having a separate shared key for broadcast haveing not occurred.
In order to make a separate key for each device viable it would be necessary to use some public key technology. But this is pretty easy, manufacturers of cable modems are already installing private keys and certificates in each device. Use of a modern PKI interface such as XKMS means that the card does not need to be at all complex.
It would be a good plan to swap out the RC4 algorithm in favor of AES. The chips in the cards are not up to 3DES at 11Mbs but they should be up to AES.
Nothing I have described cannot be implemented as an upgrade to the firmware of existing hardware. The extra lines of code would be relatively small.
In my view it has much more in common with Rambus. Dolby has been pushing their solution as a "standard" sic!.
Not really, Dolby has never concealled the fact that it is hawking a patented technology. Nor have they joined a standards group in order to extend their patents to cover the technology developed by others.
AC3 / Dolby Digital is not a simple spec. There are several patents required to implement it. The filling dates appear to be in the 1990s, the first AC3 movie came out in 1992 (Batman returns) which means that the technology should be out of patent in 2010.
ASN.1 is well known outside of the IETF fundamentalist crowd.
Always nice to start with a nice Ad Hominem jibe. I'll try one myself "ASN.1 is supported mainly by the failled has-beens who designed OSI".
With its PER (packed encoding rules), it is very efficient of bandwidth and not all that CPU intensive either.
Utterly misleading. ASN.1 encoding rules are relatively simple, the data model is the big smelly dung heap to be avoided. Although the encoding rules are 'simple' the Derranged Encoding Rules (DER) used in X.509 require multiple recursive passes through the data structure to encode it.
The only reason the Internet doesn't use it more is the usual NIH.
On the contraty, several IETF protocols have used ASN.1 and the experience has been pretty miserable. The biggest problem being that ISO keeps tweaking the spec in ways that break existing implementations. ASN.1 is simply too much of a pain in the ass for the limited advantage it provides.
The group's attempt to claim ASN.1 as the savior of HTTP is ignorant and stupid. There have been many proposals to compress HTTP headers and ASN.1 is actually one of the worst performers on both overhead and performance. The reason none of the proposals have gone anywhere is that there is no point in a backwards-incompatible change that saves 100 bytes or so on the headers if you don't do something about compressing the body. The biggest mistake we made in HTTP was not putting a simple huffman coding compression algorithm for ASCII text into the server and browsers. Actually the reason we didn't get arround to it was that nobody wanted to mess arround with the patent minefield.
Still it is always easier to explain that the reason the world is not using your idea is because they are stupid and ignorant and not because your idea is stupid and ignorant. In the case of ASN.1 the idea is a good one but the execution if third or fourth rate at best.
Re:Apple must have a brain-wash app.
on
Mac Rants
·
· Score: 2
So in fact it is entirely true that the G4 does less per instruction and more per cycle than the P4
Twaddle, at this point the Pentium is so insanely pipelined that it is executing multiple instructions per cycle.
The point about RISC is that simplifying the architecture reduces the cost of achieving a given level of performance. Development of the PowerPC chips is all but ended, Intel and AMD throw legions of designers at their chips.
It is true that the external clock frequency means little. But when comparing G4 to Pentium/AMD you are now comparing a design that was last seriously developed five years ago to designs that have development budgets in the hundreds of millions.
Its like comparing the cars on the front of the grid in formula one with a mid-grid runner in NASCAR.
Do you understand why the richest received the majority of the tax cut? Hmmm, maybe it has something to do with the fact that the richest are taxed a much higher percentage of their income.
Since my family has been in politics for five generations I can tell you with absolute certainty that the reason was that the old rich supplied the majority of Bush's campaign contributions.
Apple must have a brain-wash app.
on
Mac Rants
·
· Score: 3, Interesting
I have always thought that Apple must have a brainwashing application built into the O/S to flash subliminal messages at the user. The original article was a Troll, no mistake. But you have to admit, it was a good troll
It is not a bad thing when someone hits a group right bang smack between the eyes with the facts causing them to gibber impotently as their ideological sacred cows are slaughtered.
McWeeniedom is much like membership of Scientology they take all your money and give you something in return that only members of the cult call 'advanced technology'.
The laughable comparison chart is as ridiculous as the folk flaming "go do Comp Arc. 101 and learn about the difference between CISC and RISC". Then the curious statement is made that the G4 is faster than Pentium 4 despite the slow clock speed because the G4 RISC instructions do more per cycle. Clearly the several people who made the statement would fail their comp arch course. The RISC strategy was to reduce the complexity of individual instructions, specifically avoiding the type of complex instructions that cause pipeline faults. The other part of RISC was to simplify design to allow faster to market exploitation of the latest Fab.
In short to defend the G4 the traditional RISC/CISC argument is turned back to front. You go to RISC architecture because it allows you to push for higher clock speeds faster.
There are plenty of good benchmarks around. SPECMarks, CERN Units, MFlops, etc. and most of them are cross platform. Any benchmark that fixes on a particular piece of code that was hand coded in assembler for one platform is utterly bogus.
The biggest flaw in the article however is that the majority of the marks are given for allowing the user to select their own configuration. The whole point about the PC is that you get to choose exactly the configurtation and price point you want.
So scoring 1 point for a crappy Iomega Zip drive I would never use is beside the point. Anyone who wants to pay Iomega for their overpriced faulty trash can do so. Compact Flash is rapidly approaching the cost of ZIP disks, is smaller, more reliable and has capacities up to 1Gb.
Other folk have pointed to the bogosity of giving the Apple 2 points for L3 cach and the PCs 0.5 points for a 400MHz system bus. But the fundamental error is that the processor ratings are on the basis of benchmarks that test the whole system but are then applied to the processor alone. So on the basis that an Apple was found to be equivalent to last years model of PC on a dubious benchmark the Apple gets 12.5 points and the PCs get 10.5 or less.
The 802.11B scores are also bogus, to enable the apple you need to buy an extra card, to enable the PC you need to get an extra card. The only difference is that only Apple can supply the card for the G4.
The stupidest of all is the 'virus' line. The only reason that the Mac has not been plagued is that the population of Macs is too low to allow contagion to spread. If the Mac ever became popular again it would be slaughtered. All this reflects is the fact that virsu writers have abandoned the Mac along with almost every other software maker. If you really want to be guaranteed Virus free go run Open Genera or Multics.
Slashdot Mirror
Yes, and it would be an ilegal tied sale if Intel tried to require it.
Intel seem to have screwed up really badly here. RDRAM is going to remain expensive, not least because RAMBUS is suing all the fabs with their smurfed SDRAM 'patent'.
A billion dollars of RAMBUS stock may sound a lot but that is nothing to losing 2% of market share to AMD. It appears to me that Intel have stalled the P4 range entirely with this deal.
Plus the RAMBUS stock that Intel will get is probably worth much less than a billion dollars right now, after the SDRAM fraud finding RAMBUS has been in free fall - remember that the judge only reversed some of the fraud findings and did not reverse the non infringement ruling.
The genius that made this mess probably gets paid several million a year with millions of options on top. After this fiasco he probably wishes they were puts not calls.
Not that Netscape ever played the standards game honestly. Their idea of a standards submission was to fax a copy of the spec to MIT a few hours before they released the code (yes it was a fax, not even an email!).
Netscape spent a lot of PR bucks claiming that Marc Andressen was the inventor of the Web, they even had a PR flack write a book 'Archotects of the Web' that gave Tim Berners-Lee (the true inventor) less than four paragraphs. W3C and MIT set up Microsoft to screw Netscape in return.
By the way Sun did not 'win' the Java suit, they accepted Microsoft's offer of a settlement. Clearly Microsoft understood the consequences of the settlement better than Sun. Like be carefull what you ask for, you might get it.
Actually you can. Ever wondered why Neutrasweet is still under patent after all these years? Their lawyers bought enough congressmen to get a bill passed to extend the life of their patent.
The drug companies bought a similar extension while the GOP had the Senate. They might not have sold anything as wicked as pardons to tax dodgers but they certainly sold legislation.
Unless you are talking about one of the striped UNIX distributions with all the non-essentials taken out neither UNIX nor Windows NT is secure.
UNIX was certainly not designed to be secure as many would have people believe. Security consists of a bunch of ad-hoc add ons. Windows NT does have a security architecture in the kernel, pity that the authors of IIS had such a fondness for toxic feature sets like active code.
That is true, I got to karma cap in 50 posts and I am a veteran member of the MIT AI lab UNIX-Haters list. And that ran long before Windows NT 3.5 came out.
Actually it is hilarious to see items from the Unix haters list getting recycled as anti-Microsoft items. Security in particular, Richie and co barely had a clue and as for Eric Allman (author of Sendmail) his software still holds the all time record for CERT advisories of any application.
There is actually a significant difference between active-x and plug ins. The default security settings for the browsers are to require active-x code to be signed. Plug-ins are just another executable to download off the net and install, most are unsigned and because there is no built in download proceedure there is nothing to guide the user to require them to be signed.
Both active-x and a Netscape plug-in have unlimited ability to screw up your machine to the full extent of your user privs - which on a Windows 98 box is root. A signed active-x control does at least come with an assurance that the signer has been authenticated and has undertaken not to include malicious code and that the control has not been modified since. That is not perfect security, but it is a lot better than you get without the signature. If the signer does screw you up you have evidence that proves what he did and a contractual undertaking from him to wave about in court.
They are not. Cringely filters his entire world view through the lens of Microsoft hatred. It is like asking one of the people throwing bombs or shooting protesters in the middle east for an accurate assesment of the 'peace' process.
The Eolas patent is utterly bogus. There is racks of prior art for using embedded applets in the Web. There were embedded applets before NCSA even began work on Mosaic.
Since the Eolas case has been in the discovery stage for years it is difficult to see how an outside observer could make any accurate statement on the proceedings. The only rullings that would be heard to date would be the standard motions to dismiss that are always filled.
Fact is that Eolas probably don't want to allow the case to go to court because if they lose they lose big, their patent is busted to smithereens.
Another fact is that with the Republicans keeping Clinton's 'liberal' judges off the bench and the Democrats giving the GOP a taste of its own medicine in return the number of patent cases heard each year is actually declining slightly even though the number filled is exploding. At the rate things are going patent law will be moot because by the time a case is heard the patent will have expired. Hooray!
So Microsoft owned a lot of apple but they also owed rather a lot of apple shares. Smart move.
Many of the features look pretty sensisble. There is now pretty unanimous support for dropping Multiple inheritance. The problem with multiple inheritance being that it leads to programs only the original authors understand.
It is disapointing that the syntax was not changed more radically. I for one am pretty bored with typing semicolons at the end of lines. Using braces for block structure is equally tedious.
The garbage collector is of the 'stop everything and collect' type, this is not a good scheme as anyone who has seen a PET running Microsoft Basic GC will agree. The incremental GC in .NET is a better scheme, even if it is slower overall. But that is an implementation detail.
It would be good if people would start to look at adding support for parallel program execution. The threads programming model is very clunky and hard to use, in part because there is no means for the program to perform checks on access conflicts.
Also a persistence model should be part of any new language. The current division between programming language and database is a lot of wasted overhead.
I suggest that you take your own advice and pay attention to detail yourself. I said that 3DES is simply an extension of a broken cipher. Brute force is a perfectly respectable attack.
Or was your attempt to frame the argument that way simple dishonesty?
There have in fact been several attacks against DES that have lower complexity than brute force, however in practice the trivial parallelism and lower complexity of brute force tends to win. The fact that nobody has built a machine to implement Adi's attacks is irrelevant. The fact that the AES contenders were designed with the knowledge of Adi's recent techniques and DES was not is significant.
The key size of DES was reduced to 56 bits for a good reason, to ensure that the aparent strength of the cipher matched the actual strength. That may not be a big thing to you, in the cryptography community it is.
It is pretty easy to 'win' an argument like this on slashdot where most of the posters are like yourself journeymen at best and do not have the internal knowledge of the field. However you are going to find it much harder in the group you aspire to call your peers.
Oh, and if you think this is flaming, I suggest you get on the wrong end of an argument with Phil Z. or Bruce S.
Odd thing is that Microsoft is explicitly giving up control over C# in its current proposal to the ECMA wheras Sun insisted that it have sole control over the future direction of Java.
If the Open Source Community wants to have an open standards based successor to C then C# looks a better bet than Java at this point.
dotNet is not going to kill Apache for the simple reason that the concept of dotNet requires the ability to take existing applications and wrap a SOAP front end arround them so that other dotNET applications can access their resources.
In the very near future Open Source is going to become a serious threat to Sun's survival. Already Linux is taking serious market share from Sun for the same reason that Office Max furniture is taking market share from Herman Miller's Aeron chair. VC capital is harder to come by and much more impressed by a rack of cheap Linux boxes than a high end Solaris server.
For Sun to survive they have to make sure that their platform is too complex and rapidly changing for the Open Source movement to keep up while all the time pretending that they are on their side.
You are wrong, DES has been broken repeatedly. The world record for breaking DES is held by John Gilmor and Paul Kocher who built custom hardware to break it in less than two days. You can buy the book from Amazon. Or you can read the EFF press release
People might take you a bit more seriously as a 'Senior Security Specialist' if you bothered to read some of the basic litterature in the field before advising people with rather more knowledge, experience and standing in the field than yourself to "have your doctor up your antipsychotic medication".
3DES has not been broken to date. However it is 2^16 times less secure than other ciphers that have been equally intensively studied, execute considerably faster and require less keying material. That makes it a bad cipher according to most of the generally accepted definitions in the field - which by the way you will find in Bruce's book as well.
Unfortunately this will never be possible unless the USPTO ends its current corrupt scheme.
At present there is no means by which anyone can know what patent applications are in progress. It appears that even the modest changes under which patent applications would be published a year after filling are now 'optional'. What this means is that a corrupt applicant can easily file for a patent on somebody else's invention.
Another common trick is to extend an existing patent to include a claim on somebody else's invention that was published after the original application. Under the current rules the applicant gets the benefit of the original filling date.
Patent fraud is real and its legal. It is even encouraged by the USPTO.
A chess master might get an advantage from taking Ritalin in one constest but even that is doubtful. It is a very powerful drug that messes with several areas of the brain. If you are a chess grandmaster you almost certainly don't suffer from ADD. Like if you had ADD how did you get there? You might well suffer from dyslexia which is the problem that many Ritalin victims actually have it being easier to prescribe an addictive drug than diagnose the problem.
One of the ironic things about high school and college abuse of Ritalin is that the kids who take it so that they can concentrate hours at a stretch would almost certainly learn more if they took regular breaks. Fatigue is the brain's signal to the body that it needs a rest. Transfer from short term to long term memory appears to work best as a background task while the variables are not locked by another process. It is better to take regular five minute breaks than try to sit down for hours at a stretch.
LSD may have helped the Beatles write Sgt Pepper but I don't think anyone considers that it would have helped their creativity over the long term if they had kept taking it.
In other word you would not put Bruce on the same level as Coppersmith, Shamir, Rivest, Rogaway and so on, or as I put it not of the very front rank. Bruce is the Issac Azimov of cryptography, not its Einstein or Newton.
It is somewhat rich for Bruce to imply in 'Secrets and Lies' that he has suddenly discovered that security is about risk control not risk elimination. If he has only just realised that then he should probably give me credit for putting him straight since I pointed out precisely that point to him when we talked at RSA some years back. Not that I was the first to think of it by a long way.
Slow, yes. Susceptible to a meet-in-the-middle, no. Schneier, 12.3: "[If DES were a group], DES would be vulnerable to a meet-in-the-middle known-plaintext attack that runs in only 2**28 steps".
DES is not vulnerable to a meet in the middle attack but 3DES is in such a way that the complexity of breaking 3DES is only twice that of breaking DES, despite having three times the key length. That is what makes it a bad cipher, the fastest known attack is well short of brute forcing the keyspace.
The details of the attack are discussed in AP with respect to 2DES, to break 2DES you simply construct an in-memory table of encrypting forwards from the known plaintext (cost = O(2^56)), construct another backward from the known ciphertext (cost = O(2^56)) and look for a match (cost = O(2^56)), total cost = O(2^56). The attack can be extended to 3DES at the cost of performing two steps together, giving overall compexity O(2^112). It is a very well known result in the field and one the reason why those in the know are depricating 3DES, it is not a good cipher, it is merely an extension of a previously broken cipher.
Please point me in the direction of an implementable protocol which does provide perfect repudiability.
None gives perfect non-repudation, however PGP is designed to give pretty good PRIVACY even when the participants are pseudo-anonmous. It does not attempt to support a legal infrastructure, allow parties to place legally enforceable constraints on the liabilities they incur in authenticating a keyholder. As a result PGP is widely used amongst geeks but has a very limmited enterprise use. The vast majority of RFPs issued stipulate a PKIX conformant PKI.
I know Phil. He's one of the lowest-maintenance people I've ever met. Friendly as all get out, and patient with newbies. Would you care to enlighten me as to his ``combustible'' nature?
He has mellowed considerably since the FBI got off his case. However when the PEM vs. PGP war broke out, which is the time in question Phil was definitely of combustible nature. The FBI certainly did not help, but were certainly not the original cause.
Unfortunately rather than simply fix the parts of PEM that were monumentaly broken (the hierarchical CA system) Phil introduced competing formats all the way along the line.
There are 100 million email clients that ship with high quality crypto built in. However rather than leverage that deployed base you and the rest of the OpenPGP community spend your time explaining to people why they shouldn't use it.
It is always easier to state algorithms to steer clear of than ones to rely on. At this point IDEA is somewhat suspect, but when Applied Crypto 1 came out it was actually the best 128 bit cipher then available.
At this point most people are recommending AES (nee RIJNDAEL). The only reason to use 3DES is if you are forced to, there are still many banking applications that mandate DES. But 3DES is not a good cipher, it is slow and is subject to a meet in the middle attack that means that you do 3 times 56 bits of work to get 112 bits of security.
As far as software goes, practically all mail agents have S/MIME support built in. As far as security goes there are no serious attacks known against either S/MIME or PGP, beyond the fact that the chuckleheads in both IETF working groups flubbed the encryption of the subject line in both cases.
One problem with PGP is that it only really works well for confidentiality. It does not handle non-repudiation too well. Alice may know the message comes from Bob but proving it in court would be rather hard. Trusted Third Parties do have their uses.
The other technical problem with PGP is that it depends on the users being technically competent which most people are not.
The non-technical problem with PGP is the somewhat combustible nature of Phil Zimmerman. He is somewhat high maintenance. There is no reason why S/MIME and PGP use entirely different packaging formats except Phil's NIH policy, somewhat sad. The result being that Microsoft, Netscape, Lotus etc. implemented S/MIME and not PGP.
More recently the stale PKI/PGP debate has been rendered obsolete by technology such as XKMS which allows a client to use any PKI back end at all and not have to worry about how it works or how to configure it.
Second, the SSC would have cost a lot more than $2billion to complete in the first place. Try closer to $8billion, plus the running costs.
Third CERN have already started building the LHC which does the same science at a fraction of the cost and will certainly be completed first. The SSC had nothing to do with science, it was putting the US flag on the thing that was the whole point of the exercise
Fourth, the site chosen was a dump, a redneck dry county where the most intelligent natives are the numerous fireants.
The second set of papers have demolished the proposed fixes that nobody has implemented. I doubt that a workarround will be necessary since that set of proposals is now completely dead.
Time for the 'A-Team' to arrive and take over.
Read the paper. It does not matter how often you rekey or whether you buy the 40bit or 128 bit cards. The algorithm used is a stream cipher and will XOR your plaintext with one of 2^24 ciphertext streams that are generated from your key.
The attacker can cause the gateway to act as an oracle for any given ciphertext stream.
If you rekeyed every hour you would be safe (ish). However the WEP protocol does not support rekeying and everyone in the network has to use the same key. So you would have to update all your machines manually constantly.
Don ain't exactly Picasso, but I'll bet that his signature would fetch more than $2.56 on ebay, paticularly if it was on one of his famous hexadecimal checks...
The big problem with the 802.11b folk is that in the beginning they had no security people and now they only have a couple and won't actually let them do what needs to be done.
The original WEP protocol was secure as reviewed by the NSA, then they substituted a stream cipher for the block cipher for better performance, completely breaking the scheme. Truncated IVs are not a serious problem with DES, plenty of protocols use them. Truncating the IV utterly destroys the security of RC4.
The deeper problem is that WEP attempts to provide 'equivalent privacy' to ethernet. But a wired network does not just provide some privacy it provides authentication. The big problem with WEP 1 or 2 is that there is no way to stop a fired employee surfing from the car park.
At present the (sensible) companies that are deploying 802.11b on a large scale are wrapping IPSEC arround it.
The best way to solve the problem however is to fix the protocol itself, and use a different key for each card instead of the same key for every card in the network. The 802.11b chumps keep rejecting this idea because it prevents the use of broadcast - the idea of having a separate shared key for broadcast haveing not occurred.
In order to make a separate key for each device viable it would be necessary to use some public key technology. But this is pretty easy, manufacturers of cable modems are already installing private keys and certificates in each device. Use of a modern PKI interface such as XKMS means that the card does not need to be at all complex.
It would be a good plan to swap out the RC4 algorithm in favor of AES. The chips in the cards are not up to 3DES at 11Mbs but they should be up to AES.
Nothing I have described cannot be implemented as an upgrade to the firmware of existing hardware. The extra lines of code would be relatively small.
Not really, Dolby has never concealled the fact that it is hawking a patented technology. Nor have they joined a standards group in order to extend their patents to cover the technology developed by others.
AC3 / Dolby Digital is not a simple spec. There are several patents required to implement it. The filling dates appear to be in the 1990s, the first AC3 movie came out in 1992 (Batman returns) which means that the technology should be out of patent in 2010.
Always nice to start with a nice Ad Hominem jibe. I'll try one myself "ASN.1 is supported mainly by the failled has-beens who designed OSI".
With its PER (packed encoding rules), it is very efficient of bandwidth and not all that CPU intensive either.
Utterly misleading. ASN.1 encoding rules are relatively simple, the data model is the big smelly dung heap to be avoided. Although the encoding rules are 'simple' the Derranged Encoding Rules (DER) used in X.509 require multiple recursive passes through the data structure to encode it.
The only reason the Internet doesn't use it more is the usual NIH.
On the contraty, several IETF protocols have used ASN.1 and the experience has been pretty miserable. The biggest problem being that ISO keeps tweaking the spec in ways that break existing implementations. ASN.1 is simply too much of a pain in the ass for the limited advantage it provides.
The group's attempt to claim ASN.1 as the savior of HTTP is ignorant and stupid. There have been many proposals to compress HTTP headers and ASN.1 is actually one of the worst performers on both overhead and performance. The reason none of the proposals have gone anywhere is that there is no point in a backwards-incompatible change that saves 100 bytes or so on the headers if you don't do something about compressing the body. The biggest mistake we made in HTTP was not putting a simple huffman coding compression algorithm for ASCII text into the server and browsers. Actually the reason we didn't get arround to it was that nobody wanted to mess arround with the patent minefield.
Still it is always easier to explain that the reason the world is not using your idea is because they are stupid and ignorant and not because your idea is stupid and ignorant. In the case of ASN.1 the idea is a good one but the execution if third or fourth rate at best.
Twaddle, at this point the Pentium is so insanely pipelined that it is executing multiple instructions per cycle.
The point about RISC is that simplifying the architecture reduces the cost of achieving a given level of performance. Development of the PowerPC chips is all but ended, Intel and AMD throw legions of designers at their chips.
It is true that the external clock frequency means little. But when comparing G4 to Pentium/AMD you are now comparing a design that was last seriously developed five years ago to designs that have development budgets in the hundreds of millions.
Its like comparing the cars on the front of the grid in formula one with a mid-grid runner in NASCAR.
Since my family has been in politics for five generations I can tell you with absolute certainty that the reason was that the old rich supplied the majority of Bush's campaign contributions.
It is not a bad thing when someone hits a group right bang smack between the eyes with the facts causing them to gibber impotently as their ideological sacred cows are slaughtered.
McWeeniedom is much like membership of Scientology they take all your money and give you something in return that only members of the cult call 'advanced technology'.
The laughable comparison chart is as ridiculous as the folk flaming "go do Comp Arc. 101 and learn about the difference between CISC and RISC". Then the curious statement is made that the G4 is faster than Pentium 4 despite the slow clock speed because the G4 RISC instructions do more per cycle. Clearly the several people who made the statement would fail their comp arch course. The RISC strategy was to reduce the complexity of individual instructions, specifically avoiding the type of complex instructions that cause pipeline faults. The other part of RISC was to simplify design to allow faster to market exploitation of the latest Fab.
In short to defend the G4 the traditional RISC/CISC argument is turned back to front. You go to RISC architecture because it allows you to push for higher clock speeds faster.
There are plenty of good benchmarks around. SPECMarks, CERN Units, MFlops, etc. and most of them are cross platform. Any benchmark that fixes on a particular piece of code that was hand coded in assembler for one platform is utterly bogus.
The biggest flaw in the article however is that the majority of the marks are given for allowing the user to select their own configuration. The whole point about the PC is that you get to choose exactly the configurtation and price point you want.
So scoring 1 point for a crappy Iomega Zip drive I would never use is beside the point. Anyone who wants to pay Iomega for their overpriced faulty trash can do so. Compact Flash is rapidly approaching the cost of ZIP disks, is smaller, more reliable and has capacities up to 1Gb.
Other folk have pointed to the bogosity of giving the Apple 2 points for L3 cach and the PCs 0.5 points for a 400MHz system bus. But the fundamental error is that the processor ratings are on the basis of benchmarks that test the whole system but are then applied to the processor alone. So on the basis that an Apple was found to be equivalent to last years model of PC on a dubious benchmark the Apple gets 12.5 points and the PCs get 10.5 or less.
The 802.11B scores are also bogus, to enable the apple you need to buy an extra card, to enable the PC you need to get an extra card. The only difference is that only Apple can supply the card for the G4.
The stupidest of all is the 'virus' line. The only reason that the Mac has not been plagued is that the population of Macs is too low to allow contagion to spread. If the Mac ever became popular again it would be slaughtered. All this reflects is the fact that virsu writers have abandoned the Mac along with almost every other software maker. If you really want to be guaranteed Virus free go run Open Genera or Multics.