?

Log in

No account? Create an account

Previous Entry | Next Entry

Computer Science versus Computing

Computer science is no more about computers than astronomy is about telescopes.
    -Edsger W. Dijkstra

Discuss.

--
Banazir

Comments

( 24 comments — Leave a comment )
zengeneral
Jul. 6th, 2005 08:41 am (UTC)
only because I am highly caffeinated!
I combat your famous quote with my soon to be famous quote:
"Telescopes extend our knowledge in astronomy in the same way that computers extend our knowledge of computation, further" - Barber

How would modern day astronomy cope with the telescopes of Galileo? How will Computer Science change when quantum computers are household appliances? All science is done in the mind; our tools provide us with experience to extend science by letting our mind see the whole picture. So, to offer up some flame-bait, Dijkstra wasn't correct. Of course, Dijkstra wasn't wrong either.



All progress := (idea, use it, generalize, produce a tool)*

zerovector
Jul. 6th, 2005 05:50 pm (UTC)
Re: only because I am highly caffeinated!
I love how cocky you are.
zengeneral
Jul. 6th, 2005 06:02 pm (UTC)
Thanks
and I love that you love how cocky I am, :)
banazir
Jul. 6th, 2005 07:03 pm (UTC)
OMG
OMG, get a room!

--
Banazir
hermes_imagod
Jul. 7th, 2005 03:01 am (UTC)
Re: only because I am highly caffeinated!
(Anonymous)
Jul. 6th, 2005 01:16 pm (UTC)
I'd like to see astronomy without telescopes.
banazir
Jul. 6th, 2005 04:26 pm (UTC)
Music without instruments
You might well ask to see music without instruments, or even without instruments and voices. It still exists; we just have some practical limitations in exploring it.

The sea is still there, and crossing it to get to other continents, without sailing ships. For many centuries, mankind needed them to cross the sea, but the continents are now in reach without sailing ships. Perhaps the stars will be in reach without telescopes, or even starships, someday.

The question Dijkstra's remark really poses in the listener's mind is: what is the real stuff of computer science? It isn't computers, it isn't just artifacts (architecture or programming). It is the science of information and its representation, and of the useful manipulation thereof: automated reasoning, search, learning, transformations, and discovery. Everything from comptranslation to distributed operating systems to applied computational science and engineering (simulation in computational physics, bioinformatics, CAD/CAE/CAM) rests on these.

I do qualify my list with "automated" in that CS is not (yet?) a pure "information science", at which point it would become a subfield or sister field of mathematics. However, the model of computation is part of the science itself: our study is not limited to von Neumann architectures, conventional plus quantum computers and information systems, or even Turing machines. Hence, a computer scientist, like an applied mathematician, is a different animal from an actuary or a computer engineer. We often remark in engineering that chemical engineering and chemistry are entirely different fields; this is also true for ECE and CS.

Taken to an extreme, we might consider Michaelangelo's comment that the statue was there in the block of marble, and he was only liberating it. That's why I always talk about discovering an algorithm rather than developing or inventing one. In that regard, an algorithm is more like an equation than like a program.

--
Banazir
zengeneral
Jul. 6th, 2005 06:21 pm (UTC)
Re: Music without instruments
what is the real stuff of computer science? It isn't computers
"real" stuff? so, are computers not real? are they fake?

As you know, I have been writing a book (yes, this means that anything I say correct), and I address this issue in the first chapter. Dijkstra wasn't correct, Computer Science is about computers, the mistake is assuming computer science is “exclusively” about computers or “science”. Inferring that computer science has a purity about it is just plain wrong, computer science creates and manipulates computational models which are assessed and implemented by computer engineers. Computer Science experiments with a computer to produce segments of new ideas, new problems, new structures, new representations, new information models, new algorithms. Science := (experiment, theory(, theorem))*

Computer Science is a bastard child field of Mathematics (other parent is computer engineering). Is there a problem with being a member of a bastard child field? no... unless you are an elitist, :)


(Deleted comment)
banazir
Jul. 7th, 2005 02:24 am (UTC)
Chipping off the bits you don't need
Awesome.

And applicable in some fields of CS, too. Though we talk about semi-decidability and recursively enumerable but not recursive languages (those for which a Church-Turing computational model can answer "yes" when a string is a member but cannot necessarily answer "no" when it is not), the real application of computability theory is through specific instances and finite groups of instances.

Besides: even when there are an transfinite number of bits you don't need, you can still chip.

--
Banazir
(Anonymous)
Jul. 8th, 2005 08:56 pm (UTC)
Re: Music without instruments
You might well ask to see music without instruments, or even without instruments and voices. It still exists; we just have some practical limitations in exploring it.

Exactly. There's music in everyday things - the thing I notice the most is the two-note chord in the train whistles on the prairies.
poovanna
Jul. 6th, 2005 03:44 pm (UTC)
I think Dijkstra's statement is a warning that getting too attached to a tool (conventional computers) might cause us to miss other potentially interesting stuff (eg., non-silicon based computing like bacterial computation or quantum computation). It's important to maintain a degree of independence from our tools. Theory helps in that regard.
banazir
Jul. 6th, 2005 04:07 pm (UTC)
Independence from the tools
Precisely. Equally, I think astronomy exists to study an entity - the cosmos. Planets, stars and space, interstellar and intergalactic, would still exist without the now-critical tools, and they will still exist when we have much better tools for studying them. Similarly, data, information, and knowledge will still exist.
"Die ganzen Zahlen hat der liebe Gott gemacht, alles andere ist Menschenwerk." (God made the integers; all else is the work of Man.)
    -Kronecker


--
Banazir
prolog
Jul. 6th, 2005 03:53 pm (UTC)
Sure. Computer Science is about what we can do with computation, from the abstract to the practical.
banazir
Jul. 8th, 2005 04:40 am (UTC)
Indeed
And for the record, I agree entirely with Dijkstra, though not necessarily for the reasons everyone expects. My comments above and below clarify some of my own position, and I hope to post a follow-up soon.

There are a lot more quotes of this variety, too!
One I read by Scoville and other by Gilmore are particularly relevant.

--
Banazir
f00dave
Jul. 6th, 2005 03:58 pm (UTC)
"Logic doesn't apply to the real world." -- Marvin Minsky

My AI/PR/ML research group has concluded that the current theory of computation (i.e. the Turing Machine) can never achieve anything resembling any reasonable level of intelligence, save as tediously crafted and highly brittle custom solutions: this is the opposite of human intelligence (or of mammalian intelligence, or of even more primitive levels). The Universe is structural in nature, and the Turing Machine is based on natural numbers and logic, the former being insufficiently structural (only one operation: successor [see the Peano axioms]) and the latter being a contrived language that doesn't reflect reality (predicates are insufficient to reflect the richness of reality). Generative grammar-based systems are closer than anything else, but also have their own flaws, namely ambiguity when it comes to extracting the grammar used to generate a training set.

Indeed, our group suggests that the problem boils down to one of representation. Information is currently represented (in computers) either using a vector space of some dimensionality (measurements, images, etc), strings (which lack a formative history, and are thus ambiguous), predicates (which anyone who looks into Expert Systems or Inferrence Engines can rapidly see are totally brittle [Cyc cannot work! :-D ]). These representations all have their strengths, sure, but their weaknesses destroy them, even in hybrid systems.

Boiling the problem of representation down even further, we arrive at the following point: a training set, drawn from a class, should have the corresponding class representation expressed in the same language. This isn't true of any of the above representational systems, and is why they're all doomed to failure (among other things =] ). We -- white angels of grace that we are (*smirk*) -- have a solution: ETS! It's long, it's technical, and it's pretty damned new. But we really think we're on to something hot, here. In any case, here's a link to the main page for our group, and here's a link to a (preprint) of our latest (SEXY!!!) paper.

Sorry for barging in on your thread like this, Bana, but ... but ... I CAN'T HELP IT! =D
zerovector
Jul. 6th, 2005 05:48 pm (UTC)
That is a good one.
seldearslj
Jul. 7th, 2005 01:47 am (UTC)
Computer science = field. Computers = tools.
Astronomy = field. Telescopes = tools.

So, yes, he's correct. The one is not the other, but the two are almost inextricably related.
banazir
Jul. 7th, 2005 02:16 am (UTC)
On "inextricable" relations
Tools, yes. Distinct, yes.

I'm going to be the Inigo Montoya to your "inextricable", though. While the practice is incomplete without the tool, the theory remains quite intact. So what I think Dijkstra is driving home using the word "about" is that the object of CS is the mathematical foundation of computing rather than the artifacts and instruments thereof. A subtle distinction when it comes down to it, but one I think has to be made.

Nice icon, BTW.

--
Banazir
seldearslj
Jul. 7th, 2005 02:28 am (UTC)
Re: On "inextricable" relations
Thus the 'almost' before the 'inextracable'. ;)

But point taken.
seldearslj
Jul. 7th, 2005 02:29 am (UTC)
Re: On "inextricable" relations
Inextricable. Inextricable.

BAH!
banazir
Jul. 7th, 2005 02:40 am (UTC)
InexTRICable!
"You say that word a lot..." ;-)

I need an Inigo Montoya icon! "My name is Inigo Montoya. You overloaded my Opteron cluster..."
Or: "My name is Banazir Galbasi. You crashed my IMAP server..."

--
Banazir
seldearslj
Jul. 7th, 2005 02:44 am (UTC)
Re: InexTRICable!
"You say that word a lot..." ;-)

Yes, but I know it means what I think it means!
banazir
Jul. 8th, 2005 04:40 am (UTC)
Re: InexTRICable!
Hee hee hee...

I really do need an Inigo Montoya icon! :-D

--
Banazir
( 24 comments — Leave a comment )

Latest Month

December 2008
S M T W T F S
 123456
78910111213
14151617181920
21222324252627
28293031   

KSU Genetic and Evolutionary Computation (GEC) Lab

Teunciness

Breakfast

Science, Technology, Engineering, Math (STEM) Communities

Fresh Pages

Tags

Powered by LiveJournal.com
Designed by Naoto Kishi