Slashdot Mirror

1+ for easily and percisely tag almost all photos we were able to stuff in it. In microseconds.
This tech is old for the 2d face work. Its fast for local police Privacy concerns? UK police test 'faster-than-ever' facial recognition software http://rt.com/uk/173292-facial... (July 16, 2014)
Or just read the public info on records per second in the 10,000 records/sec http://www.nec.com/en/global/r...

From the announcement in the quoted article:
"A consortium of six global companies announced that they have signed commercial agreements to build and operate a new Trans-Pacific cable system to be called “FASTER” (...) The six-company consortium is comprised of China Mobile International, China Telecom Global, Global Transit, Google, KDDI and SingTel."
The OP gives the wrong idea that Google backs up the project and the others are involved only in management, which seems incorrect from the original announcement in NEC's page.

ADS/CPDLC runs over ACARS and definitely sends position data. Believe me, I used to code this stuff for a living.

What this guy says. I also use ACARS on a daily basis. We don't have ADS-B, and our ACARS still sends position info.

ADS/CPDLC runs over ACARS and definitely sends position data. Believe me, I used to code this stuff for a living.

IBM don't make PCs any more

That's funny, because I see them selling workstations and tower servers. Those are PCs.

Dell started 8 years after Apple (and after the Mac)

Dell traces its origins to 1984, when Michael Dell created PC's Limited while a student of the University of Texas at Austin. The dorm-room headquartered company sold IBM PC-compatible computers built from stock components.

Acer started 13 years after Apple (and after the Mac)

The Micro-Professor MPF-I, introduced in 1981 by Multitech (which, in 1987, changed its name to Acer)

NEC don't make PCs any more

Really?

Sony made their first PC 7 years after Apple (and after the Mac)

Really?

Cray never made PCs

"Every company that made computers when we started the Mac, they're all gone"

Fujitsu only started making PCs 14 years after Apple (and after the Mac)

In 1954, Fujitsu manufactured Japan's first computer, the FACOM 100, and in 1961 launched the transistorized FACOM 222.

FM-7

Been done already, way back in 2004 actually...
http://www.nec.com/global/features/index18/index.h tml

NEC sells an enterprise mobile phone solution that uses wifi in the office and UMTS outside.

Probably all working for the Chinese

Alcatel: The parent company is based out of France, with close ties to the government. Probably pro-open source.

Ericsson: Sony owns them. This won't last. Sure, they've got a good track record, but...

Motorola: they're in it to make money, acquiring open source companies and selling linux-based phones.

NEC: They jumped on Itanium for their cluster platform, so they joined OSDL two years ago, probably to make sure their investment paid off.

Siemens: Just barely joined the OSDL. Siemens Communications is primarily a hardware company; from my POV they're just trying to push their profit margin.

Nokia: they seem pretty secure as a cell phone company; I think they're into OSS genuinely to benefit the community. Take a look at what they're Open Sourcing.

Their contributions to open source notwithstanding, it looks like they want to:

1. Form alliance, apply magic words "Open Source"
2. Post article on slashdot, improve public image
3. Wait for OSS community to write their software
4. Sell COTS hardware to upgrade cell networks
5. Profit!

Of course, maybe they're working on Carrier Grade Linux just so they don't have to buy Micro$oft products any more.

I too have followed the putty-as-socks-proxy route described by others.

For enabling stuff like iTunes, which doesn't know from SOCKS, try SocksCap.

Finally, I used to have a filter on my work machine's Outlook that would run a program when a message with a particular subject came in from me--the program would ssh-tunnel back to my home machine thus enabling me to log in to work from home, but also establishing the connection only when I wanted it.

"For those with a short memory, NEC is a "baby bell" (an AT&T spin-off)."

BZZT! NEC http://www.nec.com/ is Nippon Electric Corporation, an immense japanese conglomerate founded in 1899. You're probably thinking of NCR, which was swallowed, never digested, and subsequently regurgitated by AT&T.

Your third real option is probably Fujitsu or IBM. The issue isn't just the interconnects, as you can buy those from Myricom http://www.myri.com/ or Quadrics.http://www.quadrics.com/ PNNL did this for their monstrous Itanium-2 system. It's also memory bandwidth, disk throughput, and that some jobs really require vector processors.

What Microsoft brings to the table, beyond incompatibility, overhead, and confusion, is really beyond me at the moment. They have a possibility of making an impact on high-throughput systems, but I can't see what they offer high-performance users.

People should have been using FBSD for wifi access points and aggregation long ago. Glad to see the corporate folks have finally figured out how linux errodes the bottom line.

The
stupider
ones
haven't
figured
it
out
yet.
You'd
better
go
tell
them!

Google Scholar - fast, link to articles that cite it, no reference links, multiple sources, instituitional access, reliable, good ordering

CiteSeer - slow, citations, references, multiple sources (including local), bibtex entry, not a reliable server

PubMed - monster of medical related, fast, no citations, no references, single source, fast, reliable, not best ordering

Scirus - fast, reliable, no citations, no references, single source

Best feature wise is Citeseer but for overall experience Google Scholar puts on a good show

FreeBSD can run on enterprise class hardware if you feel like buying enterprise class hardware.

No, actually it can't.

This, this, this, this, this, this, even this, and of course this is enterprise class hardware. FreeBSD runs none of them, so what enterprise class hardware *does* it run on?

FreeBSD doesn't even understand NUMA, which is basically indispensable for Opteron, POWER, or any serious Intel IA64 or x86 platforms.

• The Complexity of Sliding Block Puzzles and Plank Puzzles
• The Nondeterministic Constraint Logic Model of Computation
• Generalized Rush Hour is PSPACE-complete
• Computational Complexity of Games and Puzzles

Before, I had to write a meta-search engine for research papers (papersearch), but maybe now all we'll need is google.

Well if it is the fastest then what is this.. http://www.hpce.nec.com/445+M5043fbf9d60.0.html/ the NEC supercomputer does 65TFLOPS..

There is no mystery about "porting" GFS to SX, because it runs on a TX-7 Itanium Linux server, as a dedicated SX peripheral.

Here is another example

High-speed File Sharing Through GFS: GFS (Global File System) is a file system that enables sharing between several nodes within a SAN (Storage Area Network) environment built on Fibre Channel. By using the TX7 series GFS, NEC offers high-speed file sharing with the SX supercomputer series.

Imagine a 16-processor TX-7 Itanium GFS server dedicated to serving GFS for a 128-processor SX system.

Most people are overwhelmed just by the thought of a 16-processor TX-7 by itself. Now imagine the TX-7 being dedicated solely for use as a file server for a SX supercomputer. Separate machines, separate OSes.

Some customers who buy such hybrid SX/TX-7 systems want to use their TX-7 nodes for computation and not just I/O, and so NEC sells software development tools for Itanium systems as well. But the competition in Itanium server space is fierce, with SGI, HP, et.al. having similar offerings.

They couldn't call it the Nintendo Entertainment System, because that was taken (NES), and definitely wouldn't do the system justice. They thought they were safe with Nintendo Entertainment Console - but that's taken too (NEC).

You might be refering to citeseer, which lists (and caches) papers from websites and allows you to see which papers cite it and which papers it cites and how similar it is to other papers in the collection :-)

Ian

And even earlier, as freshmeat points out:

Paul Graham kicked off a flood of mail filters implementing Bayesian filtering with his "A Plan for Spam" article in August 2002, though it was far from a new concept. In fact, ifile has used a Naive Bayes classification algorithm since August 1996 to automatically file mail into folders. In academic circles, Bayesian methods have been used in text classification for many years, and for spam detection prior to Graham, as evidenced by the 1998 workshop paper A Bayesian Approach to Filtering Junk E-Mail by Sahami, et al.

Don't be silly, Apple is a very small fish in a very big pond compared to real supercomputer companies like SGI.

What the big companies do is let you login to their machines free of charge, over the internet, and take their machines for a spin. See e.g. http://www.testdrive.compaq.com

On top of that, what the companies with a *serious interest* in high performance computing (like SGI and Cray) do is let you email your program to them: they will spend days or weeks (up to you) tuning it, telling you how it works on their hardware, and even telling you that you're better off with the competition, if that ends up being the case. What these companies understand very well, that Apple doesn't, is that people who spend $10M+ on computing equipment, typically with public money, need to show due diligence in their choice of hardware. This means THOROUGH evaluation: benchmarks, benchmarks, benchmarks - and not _standard_ benchmarks, but the benchmarks that matter: the user's own software.

Take a look here to see a real-life snapshot of this kind of process.

Any company that tried to "fudge the numbers" would be caught out, and that looks VERY bad. So the companies instead do all they can to help with the evaluation process, and hope that they get chosen. If not, there's always the next sale. A big supercomputer is sold somewhere every day or two, after all.

Virginia Tech's "X" doesn't come into this category: for example, no actual scientific work has been performed on the machine so far apart from benchmarks/system development, etc. Dr Varadarajan is probably going to get a pretty nasty grilling in a couple of years, when the University asks "so, what did we get for our money?" Some P.R., and not a whole lot more it would seem. I wish them luck...

Apple is not really a supercomputer company: and really, who cares? Apple has been, is, and will probably remain a great computer company for a good long while yet. But don't expect Apple to help you out a great deal if you pop them an email going "say, how does the G5 run on the ASCI Purple benchmarks?". That's just not what Apple are about. (example: they are happy to sell you a box with a 64-bit CPU, but they could care less that they don't have a 64-bit OS to go with it..)

You can often do better than Google's web search for specialized searches. Froogle or Yahoo! Shopping works better for finding products. Google News works better for finding news articles. IMDB is better for movie reviews and information. CiteSeer is better for finding CS research papers.

While someone may eventually beat Google for general web search, it's these niche searches that offer a lot of easy opportunity. Because Google is a general purpose search engine, it's not too hard to beat it in a specific topic areas.

Now that I have your attention, I'll let you in on a little secret: to get tenure at MIT you have to have a ton of publications and citations. There ain't no other way, no matter now much money you raise. The esteemed Professor Hawley is a case in point: raised millions, but was denied tenure. On the other hand, some folks you might not have heard of, such as Pattie Maes, Justine Cassell, Ted, Adelson, Neil Gershenfeld, Steve Benton, Sandy Pentland, Roz Picard, and others I can't think of at the moment, have earned MIT tenure, published influential work, and graduated PhD students many of whom teach and train PhDs at well known schools.

Disclaimer: I got my PhD at the Media Lab, but I don't teach or train PhDs at any well-known school.

Now that I have your attention, I'll let you in on a little secret: to get tenure at MIT you have to have a ton of publications and citations. There ain't no other way, no matter now much money you raise. The esteemed Professor Hawley is a case in point: raised millions, but was denied tenure. On the other hand, some folks you might not have heard of, such as Pattie Maes, Justine Cassell, Ted, Adelson, Neil Gershenfeld, Steve Benton, Sandy Pentland, Roz Picard, and others I can't think of at the moment, have earned MIT tenure, published influential work, and graduated PhD students many of whom teach and train PhDs at well known schools.

Disclaimer: I got my PhD at the Media Lab, but I don't teach or train PhDs at any well-known school.

IIRC, the Newton has one truly custom chip (which does memory interfacing, DMA, interrupts, timers, A/D for the tablet and so forth).

The kernel makes extensive use of the ARM 610's MMU (especially its domain and sub-page-granularity protection features), so porting the OS to another platform would be quite exciting, but the application (Newtonscript) world is pretty isolated from the wacky stuff going on the OS. You could probably fake-out a fair amount of the OS and the apps would run.

Here's a reference to a paper on the Newton OS that we presented at CompCon in 1994.

There have been a number of recent advances in concept-to-speech synthesis, which incorporates a sort of map (I'm being very general here) of the semantic concepts in a text and uses that to determine where emphasis should go to make speech sound more natural.

I think there's a lot of promise in fusing this approach with the discourse-representation work of people like Daniel Marcu -- automatically extract the discourse representation, then use that to assign prosody.

Sorry if this is a bit technical -- I do something very similar to it for a living, so it's easy to geek over.

"Stacking" isn't as good as full Bayesian image super-resolution when systematic noise is present; see the theoretical discussion in section 7.5 of Bretthorst's book (he refers to stacking as "averaging"). For a Bayesian method, see this paper, and some of these references.

The idea was that you could take a standard resolution video that panned across a scene, and by merging the frames over time create amazingly high resolution images.

The idea was that you could take a standard resolution video that panned across a scene, and by merging the frames over time create amazingly high resolution images.

The idea was that you could take a standard resolution video that panned across a scene, and by merging the frames over time create amazingly high resolution images.

The weakest link in any cryptography protocol is the key, reguardless of how big it is.

That's totally incorrect. In fact, the weakest link is almost never the key size. The failure point in protocols often has nothing to do with the security of the underlying cryptographic primative. PGP's file format lead to a break in the PGP system without attacking any of the cryptographic primatives. Similarly, SSL was broken on netscape because of a faulty choice of random number generator. Wep was broken in 802.11b devices due a series of flaws that didn't break the underlying algorithms. Cipher design is very difficult but it's only the easy part. Protocols are a nightmare.

During the cold war some soviet spy's would use an encryption scheme where a single bit of the key would decrypt a single bit of the message, after decryption the bits of the key that were used to decrypt were thrown away. The key had to be huge and it could only be used for a certain number of messages. That type of encryption is called a one time pad, it's nearly impossible to break. The common encryption schemes today like RSA or DES go for reusable keys but you still need to switch your key's every so often.

It is unbreakable but people often sing it's praises and neglect the fact you have to get this key to the other party you want to communicate with. Since the One time pad (OTP) is the *same* size as the plain-text you want to communicate then surely it's just as easy to communicate the actual message? In some cases it's useful, like where you can give someone they pad before they start their covert mission but in a modern internet setting the OTP is useless.

Generally the idea is to make the key as large as possible. There will always be a cap in how large one can go. Limitations in computing power can make the time needed to decrypt a message with a large key unacceptable. Maybe the key needs to fit onto some ealy concealable physical medium, or maybe it needs to be remembered. The idea is to acertain your upper limit and use keys that are that length

Again, another misconception. Bigger keys do not have a huge impact on performance if your using a block cipher. I could make DES use 768-bit keys if I dumped the key schedule and used independant subkeys. It wouldn't improve it's security either. Infact, differential cryptanalysis of DES and it's varient is generally done by assuming the subkeys are independant. Bigger keys do not always equal bigger security.

I stand by my original analysis. Smaller keys are better because they're easier to protect. The only need to be big enough to resist brute-force and there is no use in increasing the size further.

If you are interested in the research behind all this, see Chenxi Wang's dissertation. Here's a paper on it. The approach recognizes the fact that security is about raising the bar high enough to make it too much work for hackers. By changing the code on the fly, the hackers have to start reverse engineering it all over again.

Deobfuscation is in NP. That is, for any type of obfuscation, there is a method to reduce it to a deobfuscated copy. It may take polynomial time, but it can be done.

Read the paper if ya don't believe me.

I wonder if they've seen the proof of the impossibility of obfuscating programs?

Also, what is the basis of a search engine?

Well, the paper from 98 that describes the PageRank algorithm (as used by Google) can be found here

Theres a simple explanation of various indexing/ranking schemes here, but if you really want to get up to speed on research into searching the web, try looking at some of the papers from the TREC Web Track

Happy reading,

Dave

> Caching is controlled completely by the CPU, transparent of the programmer.

That's not true.

> In other words, just trust the CPU. It knows what it's doing :).

Where do you get this from?

Let's say I'm a programmer, and I want to multiply two matrices. If I use the straightforward iterative approach, then this algorithm incurs many cache misses. (I'll leave this as an exercise to the reader, but here's a hint: Choose a cache size n and then pick a matrix dimension x, x > n. Then, try and figure out how many times we'll need to evict data from the cache in order to perform the multiplication and addition operations.)

However, if I use divide each matrix into four pieces, and multiply the pieces recursively, then I can minimize the number of cache misses. Plus, I don't even have to have any algorithm parameters that depend on the cache. The recursive algorithm for multiplying matrices is a cache-oblivious algorithm:

"An algorithm is cache oblivious if no program variables dependent on hardware configuration parameters, such as cache size and cache-line length need to be tuned to minimize the number of cache misses."

So, no, I won't just trust the CPU, cuz it doesn't always know what it's doing. What was your point again?

Combine this device (with a few upgrades) with Wearable displays + This Jacket (maybe an upgrade or two + a computer that speaks sign language + Pervasive wireless broadband and I am starting to get to my "comfort level" for internet access. No longer will I have to wait during my whole commute of 5 minutes to check for the next Slashdot story. No more shaking internet withdrawl on the bus!!

That's quite amazing. Now we need a computer to write music and songs.

It's been done. Bach was an early target. Heck, I was writing melody generators and harmonizers 8 years ago (badly, but I was doing it).

There are many more problems too. A good article on the problems with patents, the unworkable solutions and possible solutions can be found in Jeffrey D. Ullman's article Ordinary Skill in the Art

The "wake-sleep" algorithm for unsupervised neural networks.

Hinton GE, Dayan P, Frey BJ, Neal RM.

Department of Computer Science, University of Toronto, Ontario, Canada.

An unsupervised learning algorithm for a multilayer network of stochastic neurons is described. Bottom-up "recognition" connections convert the input into representations in successive hidden layers, and top-down "generative" connections reconstruct the representation in one layer from the representation in the layer above. In the "wake" phase, neurons are driven by recognition connections, and generative connections are adapted to increase the probability that they would reconstruct the correct activity vector in the layer below. In the "sleep" phase, neurons are driven by generative connections, and recognition connections are adapted to increase the probability that they would produce the correct activity vector in the layer above.

Don't confuse this Dcube with NEC's D-Cube .

Nothing a quick search engine couldn't solve...

Of course, there are also hard real-time garbage collectors (ie Cheng's), though I don't think you'll find them in general-purpose production compilers. However, you will find good garbage collectors in a number of real production compilers (say, in mlton). It's definitely worth benchmarking.

If we are talking about device driver comparsion, here is an interesting paper Linux Device Driver Emulation in Mach describing how Mach can use Linux's device driver without changing the device driver code. Mach which powers the Mac OS X is a very flexible micro-kernel OS. A lot of neat trick can be done with it. I wonder if there is an effort in Darwin to bring this enmulation to Darwin.

I agree with you without a doubt. Although I read the other day that, in 2003, it was estimated that we generated 2 exabytes of data in digital form. That's 2 x 10^18 bytes. 2^64 = 1.844 x 10^19, and with 16 bytes per hash, that's 2.95 x 10^20 bytes required for 2^64 different MD5 hashes. So I guess the storage for this project may exist in the future.

But I'm just being pedantic. In all honesty, who the hell cares? Brute force isn't that exciting, but some mathematical analysis may produce results, see here.

Real men use SHA anyways ;)

Cheers,

shadowmatter

Something like CiteSeer?

big citation search engine