cue

X-rated versions of these which "pulsate" when watching .......

Shouldn't that be

Al Arsi, the arsassin?

@Phil Rigby: Why only war?

- Assuming that we can get a fool-proof system, why is the first practical use have to be for a war/military solution?

Answer: only fools make war

@The Other Steve; @Brian Miller

@The Other Steve: Nice one, Dijkstra on the other hand got angry when you suggested testing something he had proved.

@Brian Miller: Not a flaw at all. It is possible to write secure C, just as it is possible to write secure assembly (after all, any secure code must be secure machine code at the final crunch). So long as you have a secure C compiler (mathematically correct) you are OK. Exactly because C is so small, as a language, it is less hard to prove correct.

@Keith T; @AC: A few replies; + "Authentication + , Identification -"

@Keith T : see my earlier comment on why this does not work, also the AC mentioned below

@AC: A few replies: No computer scientist dealing with computer vision or other pattern recognition tasks would think the way you suggest computer scientists think.

Finally: I do think biometircs, in particular iris scans can help in authentication, though more research is needed to reduce the failures in enrolement. I remember John Daugman complaining he had studied more eyelashes (especially Asian ones, which do not curl up as much, and therefore occlude the iris) than anybody in the world.

In identification I have severe doubts

@Too much egg: I actually do pattern recognition

I am afraid Julian might do stuff, but he does not understand how matching works. Almost all biometrics encode a series of spatial relationships between features. These relationships are rarely completely rotation or scale invariant (i.e. they depend on pose, camera angle etc.), therefore, you cannot map each set of parameters to a single number which you then sort. Instead, you need to match the multivariate data using some distance measure. If the distance is too large, we have a mismatch.

Iris patterns are a good example. Rather than coding to a unique number, you need to verify that the hamming distance between the two sets of bit patterns fails a test for statistical independence. Because the iris might be recorded in a slightly different orientation each time, you need to compute the hamming distances for a series of shifted combinations, and take the smallest distance to indicate the correlation.

Thus 1.8 x 10^15 is correct.

I agree with the author: the scheme is doomed to failure

There is an explanation for the politicians' and civil servants' behaviour: the usual inability for managerial types to admit they are WRONG. They consider every u-turn as loss of face. Difficult for good scientists to fathom (hallmark of bad scientist: not admitting you are wrong!), but true nonetheless.

clouds more likely to be troublesome

over here in the Netherlands as well

bummer

We'll have the beer nonetheless

WHY does the USPTO ....

award patents for inventions for which prior art so clearly exists. No formatting is lost when saving Open Office documents in XML? Wy not patent the concept of a generic markup language ((La)TeX anyone?)

In related news , the USPTO has allowed Microsoft to patent the screwdriver

I want neither my head nor data in the cloud(s)

In a vain attempt to stem the tide of mutual OS/App-bashing, the following thought. Why cloud computing is a problem is more the issue than the underlying OS:

Regardless of who wrote the OS I use (Window XP, Debian, and OpenSuse (and still PalmOS (Garnet) on my old T3)), I want to have control over my data. Not just because I do not wish to lose (access to) them, but also because there is an inherent overhead in distributed computing due to communication overheads. Yes I can process my 1GB MRI scan on the XX teraflop machine in Amsterdam fast than on my 16 core opteron or xeon server at our institute, but GETTING the XXXX data there is so SLOW it outweighs the CPU time benefit tenfold.

And until they have a system to kick sysadmins remotely, I do so much more enjoy problems with on-site facilities than off-site ones

Let everybody use the OS they like, and the applications they like or can afford, just don't lose control of or access to your data!

Nah, back to bashing:

I do all my (science) writing in LaTeX (with emacs), so who needs {MS,Open,K}Office!

A great scoop for Terry Pratchett's ...

Harry King! (a.k.a. King of the Golden River, a.k.a. Piss Harry)

I had no problems

It downloaded at amazing speed, no hitches in set-up, appears to be much faster

NICE

No SSD in The Netherlands either

What are they thinking at Dell?

What? No Discworld Camels!!!

I think all windows machines should be named after Discworld camels. I have been known to call such machines "Bloody Stupid," "Evil Smelling Bugger," "Evil-Minded Son of a Bitch", "Evil Brother-in-law of a Jackal", and of course those perennial favourites: "You Bastard" and "You Vicious Brute". Some do need abbreviation: BS, ESB, EMSoaB, EBiloaJ, YB, and YVB. Nobody ever asks what it means.

Under linux the same machines are known as:

Gauss,

Leibnitz,

Cauchy

Riemann

etc: The earthly counterpart of Discworld camels

RUBBISH: Linguini, swimming pools, Wales are better

They should skip the metric system entirely, and go for linguini, norris, brontosaurus and doubledecker buses: THAT'S rocket science

Let me be the first

to welcome our rodent overlords

One of the few sensible points made in this discussion bogged down by screen size

regards applications. My old Tungsten T3 is still going strong (after battery transplant), and has a host of third-party apps on it. I tested the later T5 and TX briefly, but their slower processors and memory were a pain, so I stuck with the old T3. None of the products offered by Palm (or any other) seemed interesting until the Pre came along. However, by making apps development exclusive this whole old ecosystem is lost, and I will therefore NOT move to a Pre.

Sorry Palm, no dice!

African Antelopes?

not swallows?

(I am not suggesting coconuts migrate)

Novelty (Super) Nova?

couldn't resist, sorry

Mine's the one with the Celestron C8 guider in the pocket

That brings back memories!

Of course, I decided then and there to become an astronaut (ultimately studied astronomy, just three letters off)

We need a saturn 5 or LEM icon!!

@Andy ORourke

Should that not be:

Phuck oph Phorm?

Looking at most user generated content

leads to the unavoidable conclusion that intelligent life is going extinct already.

But why do members of congress complain?

they have nothing to fear: if they are innocent, they have nothing to hide!!

</sarcasm>

GOT ME!!

Cheers!!

Good they had a taser!

otherwise the complete idiot might have been shot!

Of course, she would be a Darwin award candidate then, but then you can't winn'em all, can you. Especially when you have clearly lost a significant percentage of your marbles.

I seriously have NEVER heard of a woman who "due to premenstrual stress" placed the importance of discount shoes over her children.

At least she blamed a satnav AND was using a mobile phone, so the IT angle is sorted out.

It raises a question too: Is AI needed to compensate for the abundance of natural stupidity?

How long till we get

spam mail offering use of this technique to "gain several inches"?

Surely the machine goes "Ping!"

Ig Nobel candidates, the first lot, certainly (combined medicine/economics)

But can they grow....

Zaphod's extra head too?

Does anyone care about the CPU in HPC?

I don't. All my HP apps are in a higher-level language (C, C++, and even (shudder) Fortran (77, 90) (queue flames)) All I need is a good compiler. The same holds for all HPC apps I know. What is far more important than the CPU is the interconnect architecture. Designing for shared memory or different types of distributed systems requires different coding at the abstraction level of high level languages.

Depending on the speed of your interprocess communication bandwidth (rather than interprocessor, it is NOT the same on all systems) you can afford more or less communication in the algorithm. Load balancing and task distribution depend critically on how the CPUs are connected, much less on the kind of CPU. I have code that works quite happily on my dual core laptop, quad-core desktop, a quad-core opteron, a 16 CPU MIPS system, and (previously) on our old CRAY SV1e (RIP) with 32 processors, all essentially shared memory machines. It did not perform as well on the 4 x 4 core Xeon machine due to the memory-bandwidth limitations inherent in the design, but I would love to run it on a Nehalem. Run the same code on a cluster and it would CRAWL.

I think my kkkkeyboard ssssshhhhortttttteedddd

from drooling over this device (the device, NOT the babe (honest(really!)))

I wonder what you would be saying

if they wanted to spend billions on studying NOTHING. Vacuum is very interesting if you know how to look at it closely. I would back such schemes, but then I never had much truck with beancounter attitudes.

Anyway, it beats spending money on Euro/Joint Strike/Insert favourite here/ fighters or aircraft carriers that do not seem to be fit for the wars we fight now (that should get people going).

Maybe by self-competition

they want to avoid being seen as a monopoly?

Or they want to have as many versions of W-7 as there are linux distros? In this way, windows users can experience the same futile arguments as so often heard between openSUSE, Fedora, Debian, Ubuntu and other fanboys

(I have debian at work and openSUSE at home so I can take two sides in those debates ;-) )

GRIN!!

Aliens because, well.....