Re: U.S. colleges retool programming classes

[David Farber <dave () farber net>]

Begin forwarded message:

From: "Andrew W. Donoho" <awd () DDG com>
Date: June 4, 2007 11:19:21 AM EDT
To: dave () farber net
Subject: Re: [IP] Re: U.S. colleges retool programming classes

For IP, if you wish:

On Jun 4, 2007, at 07:36, Gene Spafford <spaf () cerias purdue edu> wrote:

> And as to why we want more computer scientists -- I am looking  
> forward to the new languages, running on new operating systems,  
> using new network protocols, and supporting amazing new apps that  
> have not yet been invented.  Some of those may be crafted by  
> hobbyists, but the majority will be designed, tested and built by  
> CS grads.  If you think Windows and C++ are the pinnacle of human  
> achievement, then I guess computer science isn't going to have much  
> appeal.  But dang, you sure are going to need those debugging and  
> version control skills!!



Gene touches on two issues that I think need more examination -  
scientific invention versus field stagnation.

It is very clear that the software industry is going to hit a  
programming wall some time in the next 6 years (4 Moore's Law  
generations). Microprocessors are going to progress from 4 to 64  
processor cores. Most algorithms, other than the embarrassingly  
parallel ones, will hit this wall. This wall is not going to be  
surmounted by 'code monkeys'. To truly exploit this coming surfeit of  
processors, we will need both architectural and software invention.  
This problem will need 'classically trained' disciplined thinkers.  
While I was not trained as a computer scientist (experimental physics  
was my field), the problem is going to need folks of the caliber that  
originally defined our core architectures. (The next Johnny Von  
Neumann, I'm looking for you.) In other words, we are entering an era  
of unprecedented architectural research and invention opportunity.

Yet there is a risk, which brings me to my second point.

Business.

Business, due to the needs to support a product in the field, is  
quite conservative. Furthermore, when you have market dominating  
businesses such as both Intel and Microsoft, choosing solutions  
outside of their very narrow band of options does not make economic  
sense to most customers. (Full disclosure, I work for a competitor to  
both Intel and Microsoft.) In many ways, due to these economic  
pressures, computer science has been transformed, in the public/ 
business mind, into a 'trade school'. In physics, I used the 'english  
major' defense when asked about why I studied physics instead of  
something 'more practical'. (I.e. that physics was the 'liberal arts'  
of the scientific disciplines. It would take me through many career  
changes. And so it has.) CS majors are not allowed this answer. It is  
presumed that CS is feeding into a lucrative industry. Hence,  
students are not challenged to really think about why they are in the  
program. And pressures on faculty/departments from Regents/Trustees  
reinforce this 'CS BS mill' for industry perspective. Hence, I  
welcome the spilt in the field into CS and Computer/Software  
Engineering. While both are far from being trade schools, they serve  
different societal goals. Our society needs the CS equivalent of a  
rich theoretical/experimental scientific duality. This interplay will  
bring us the next profound advances. Computer/software practitioners,  
as opposed to scientists, should enter the necessary, honored and  
lucrative profession of engineering.

If we can make this split, then we have the tougher problem of  
getting fundamental research funded. In my day job, I am helping  
frame our development/research agenda for many-core computing.  
Recently, an RFP came out from industry for $2 million/year for 5  
years to study the many-core programming problem. More than 25  
schools applied for 1 grant. While there are more funding sources  
than this one grant, I think it speaks volumes about how much hunger  
there is in the academy to tackle tough problems and how little  
funding. (Note to the 24 schools who miss out on this grant, call me.)

I am not referring to the perennial problem of funding research but,  
rather, funding the attack on the 'big problem'. Dave Patterson has  
identified the 'big problem' of many-core computing. I think the  
research community needs to really coalesce around this problem to  
both retool itself and create the CS research agenda for the next 20  
years. This should be the equivalent to the CS community as was  
building Fermilab or the Large Hadron Collider was to the particle  
physics community. Attacking this problem will, I think, cause the  
split between computer science and engineering to really form. And it  
will open a new world of computing for all of us.

Andrew

____________________________________
Andrew W. Donoho
awd () DDG com, PGP Key ID: 0x81D0F250
+1 (512) 453-6652 (o), +1 (512) 750-7596 (m)

"To take no detours from the high road of reason and social  
responsibility."
    -- Marcus Aurelius


Physicists can muster huge resources to fund Fermilab, Brookhaven,  
etc. for the big problems of physics.


-------------------------------------------
Archives: http://v2.listbox.com/member/archive/247/=now
RSS Feed: http://v2.listbox.com/member/archive/rss/247/
Powered by Listbox: http://www.listbox.com