Creator of ALICE is dying.

[Lee_Hitsugaya]

Yes, it's true...Randy Pausch has been diagnosed with cancer of the liver and will be dead before the end of 2008...I myself am in shock and must say I'm deeply saddened by it. He did a final speech and its available here..

http://www.youtube.com/watch?v=4HqdnjgkExY

This is truely the end of a visonary and an insperation.

Leave your thoughts about this tragic development here..

[monele]

I love the whole conference. And I didn't know about the guy at all. I realize we didn't learn about *any* human person while in computer science classes ô_o... They should talk about such people... dead or alive;.. those who changed and still change things in this domain.

[PyTom]

There are some really phenomenal computer scientists out there, if you read a little.

Alan Turing, you'll find out if you look into it, apart from his contribution to computer science, is one of the major reasons why the Allies were able to win the battle of the Atlantic*. When the war was over, he was deemed a security risk because he was homosexual, and eventually comitted suicide.

Edsgar Dijkstra made contributions to just about every major area of CS. AFAIK, he's the only single person to have his death commemorated on google's home page.

The list of Turing Award award winners is a list of phenomenal men (and one woman) nobody's ever heard of, which is a shame, because their work is one of the backbones of the modern computerized society.

* I think a guide at the NSA museum explained it best one time. Cryptographers didn't win the war. That was the brave fighting men's doing. But without intelligence, they could have fought just as bravely, and lost.

[monele]

Turing I heard of with his Turing Test. Dijkstra I didn't know but now I do ^_^. But again, I never heard any of these names during my classes, bleh ^^;

[PyTom]

The Turing test was, at best, one of Turing's side projects. His big contribution is the idea of a universal Turing machine.

Basically, a Turing machine is a machine consisting of a small finite amount of memory (the state) and an infinitely long tape that can have a finite number symbols written on it. At each step, the TM reads a symbol from the tape, and based on the symbol on the tape, it can change state, write a symbol to the tape, and optional move the tape left and write.

Turing then invented the universal Turing machine, which can encode any Turing machine on the tape.

What Turing (at the same time as a guy named Church) proved was that a UTM is more powerful than any computer out there, insofar as (given enough time), it can compute the answer to any problem a real computer can.

There are some fairly cool corollaries to this, as well. One is that the number of programs a UTM can run is countably infinite, but the number of problems that can exist is uncountably infinite. So this means that there are things that it is fundamentally impossible for a computer to do. Of course, nobody has ever been able to explain one of these problems, since being able to explain it probably means that a UTM could run it.

(In a nutshell, an infinite set is countably infinite if you can map from the natural numbers to it. For example, the set of even numbers is countably infinite, since you can make the mapping: 0 -> 0, 1 -> 2, 2 -> -2, 3 -> 4, 4 -> -4, 5 -> 6, 6 -> -6, and so on. The set of integers is countably infinite, as is the set of rational numbers (fractions), but the set of real numbers isn't.)

[monele]

Mmm... does this mean these Turing machines can be programmed to program themselves? ô_o

[Doc_Odd]

Monele, yes, a Turing machine can program itself. And on the limits of the UTM, Church's Thesis says that there is no intelligible sense of "computable" on which it can mean more than "computable by a UTM," and nobody has ever found any plausible grounds for rejecting the thesis. If Church is right about his thesis, "there are problems no computer can solve" would be better rendered "there are problems that can't be solved."

Of course, you might have thought those two sentences amount to the same anyway, but some people have been known to use stuff like this in efforts to prove human thought surpasses what any computer is capable of, since computers have these limits. But this is nonsense; there's every reason to think that humans have the same limits.

[monele]

I'm really thrilled by the idea of a computer being fed his own output... But I still wonder how much "base" a human would need to program to make the computer be able to reprogram itself in a useful way.

[PyTom]

Well, what do you mean by having the computer reprogram itself? Except for some embedded systems, just about every modern computer is stored-program, which means it's capable of changing the programs in its memory.

[Jake]

I'm really thrilled by the idea of a computer being fed his own output... But I still wonder how much "base" a human would need to program to make the computer be able to reprogram itself in a useful way.

Well, a turing machine still needs to have a program that it runs on. It can have a program which is really an intepreter for the information on the tape, which might be the output of another program on another turing machine which might well be interpreting a different bit of tape... but ultimately, someone's got to provide the program in the first place, or it'll just sit there doing nothing.

In fact, of course, you've seen all this before, because any desktop PC is Turing Complete in the loose sense, as are many programs for many platforms, including Ren'Py. Technically they ought to have infinite storage, 'cause Turing's tape is infinitely long, but for all practical applications "infinitely long" can be read as "very very long".

[monele]

Mmm... so... the simple fact that I code in Ren'Py, which is interpreted by the Ren'Py program... which itself is coded in Python, which itself is interpreted as a Win32 (for my case :p) application which itself is run with instructions given to the processor...... is proof of this? These are all a bunch of Turing machines?