wu :: forums
        (http://www.ocf.berkeley.edu/~wwu/cgi-bin/yabb/YaBB.cgi)
      

        riddles >> general problem-solving / chatting / whatever >> A proposal for a more creative CS education
        
(Message started by: amichail on Nov 6th, 2006, 1:46pm)
Title: A proposal for a more creative CS education
Post by amichail on Nov 6th, 2006, 1:46pm
Open-ended programming assignments are critical for fostering creativity. Yes, there is some creativity in design/implementation, but to just leave it at that would be a grave injustice to students -- particularly ones who want to become entrepreneurs.

The problem with open-ended assignments is that they are difficult to mark. Students generally demand objective marking.

So maybe think of open-ended assignments as web startups, where web traffic is the objective measure of their success.

But as it takes time to get significant web traffic, one might have students work on their honours thesis project starting in 1st year.

Another benefit from this approach is that every such assignment is a potential startup. Indeed, students may make money from their honours thesis work (e.g., via adsense) and/or get VC funding while still at university. The mark that students would get at the end could also take such achievements into account.

I think that such an entrepreneurship program would be easy to sell. It's more creative than typical CS programs. Moreover, it doesn't matter so much what the job market is like if you plan to work for yourself.
Title: Re: A proposal for a more creative CS education
Post by towr on Nov 6th, 2006, 2:03pm

on 11/06/06 at 13:46:55, amichail wrote:
So maybe think of open-ended assignments as web startups, where web traffic is the objective measure of their success.
It's a measure of succes, but not of succes in creative programming, it'd be success in marketing.
While it does have merit, in business school, I don't think it has much in CS. Although a cooperation might be fruitfull.

If I had to make a project rated via web-popularity, I'd make a webgame with a lot of sex in it. People'd swarm to it. No fancy application showing great feats of computational engineering, no, something simple that sells.
Great software is generally a niche market. The chances you'll come up with something good and popular, like a "bittorrent" or a "firefox", are next to nil.
Title: Re: A proposal for a more creative CS education
Post by amichail on Nov 6th, 2006, 2:16pm

on 11/06/06 at 14:03:23, towr wrote:
It's a measure of succes, but not of succes in creative programming, it'd be success in marketing.
While it does have merit, in business school, I don't think it has much in CS. Although a cooperation might be fruitfull.

I doubt that marketing without a compelling app/service would be sufficient.  

Quote:
If I had to make a project rated via web-popularity, I'd make a webgame with a lot of sex in it. People'd swarm to it. No fancy application showing great feats of computational engineering, no, something simple that sells.
Great software is generally a niche market. The chances you'll come up with something good and popular, like a "bittorrent" or a "firefox", are next to nil.

The high web traffic would need to be sustained for a long period of time. If the project is trivial implementation-wise, then it will be easy for competitors to arise to take some of that traffic away.
Title: Re: A proposal for a more creative CS education
Post by Grimbal on Nov 6th, 2006, 2:58pm
A simple assignment would be "compress Hamlett".  The measure is the size of the compressed data, plus the code and any data necessary to the decompression.

I read about some engineering schools who give practical assignments to students: they are given glue and wood, they have to build a bridge that has the highest (supported weight)/(own weight) ratio.  The design is free.  To evaluate the bridge they stress the bridge until it breaks.

Else, I visited an engineering school that has an army of small robots that the students must program from scratch to leave a maze.  They learn all about interfacing software with sensors and motors.
Title: Re: A proposal for a more creative CS education
Post by amichail on Nov 6th, 2006, 3:04pm

on 11/06/06 at 14:58:13, Grimbal wrote:
A simple assignment would be "compress Hamlett".  The measure is the size of the compressed data, plus the code and any data necessary to the decompression.

I read about some engineering schools who give practical assignments to students: they are given glue and wood, they have to build a bridge that has the highest (supported weight)/(own weight) ratio.  The design is free.  To evaluate the bridge they stress the bridge until it breaks.

Else, I visited an engineering school that has an army of small robots that the students must program from scratch to leave a maze.  They learn all about interfacing software with sensors and motors.

The problem with these assignments is that they ask students to solve a well defined problem.  
Title: Re: A proposal for a more creative CS education
Post by towr on Nov 6th, 2006, 3:37pm

on 11/06/06 at 14:16:56, amichail wrote:
I doubt that marketing without a compelling app/service would be sufficient.
It's enough to get hits. But I never said there wouldn't be a compelling app. Just that there's absolutely no reason to think it would be anything that requires quality programming. Ever played webgames like Utopia, Kings of Chaos, Quantum star, etc? Quite popular, but not requiring fundamentally more than a spreadsheet.


Quote:
The high web traffic would need to be sustained for a long period of time. If the project is trivial implementation-wise, then it will be easy for competitors to arise to take some of that traffic away.
First there's the question of how long a period you'd have to look at. Then there's the question of whether it even matters that there's competitors. A webgame would still attract more visitors than a sophisticated program for the simulation of the interaction of large protein structures.
There simply very few people interested in the latter, it's an academic tool. Whereas the first is popular entertainment. We****pularity is simply no way to compare such endeavours. Popularity does not equal quality.
Except for marketing.
Title: Re: A proposal for a more creative CS education
Post by towr on Nov 6th, 2006, 3:40pm

on 11/06/06 at 15:04:21, amichail wrote:
The problem with these assignments is that they ask students to solve a well defined problem.  
My suggestion woul dbe to first let the students find and define a 'problem' well, then evaluate whether that makes a worthwhile project. And perhaps find someone qualified to judge it and guide them.
Just letting them do whatever isn't doing them any favour (procrastination in the face of not having a clear goal is mighty tempting). And grading on anything other than the quality of their result is plain insulting.
Title: Re: A proposal for a more creative CS education
Post by Whiskey Tango Foxtrot on Nov 6th, 2006, 3:45pm
As an engineering student who has experienced (and enjoyed) both of the experiments Grimbal mentioned, I feel I must defend their relative open-endedness.  Also, as a student in a school where students are allowed to make up their own majors and assignments, I feel I should mention the shortcomings of an open-ended educational system.

As I said, I enjoyed the two assignments very much, especially the way we were given freedom to work in any direction we chose.  That said, there was an end result in mind.  That is, all students were expected to end up in the same place regardless of the path taken.  This is actually a contradiction of the term "open-ended," but I think we all understand what Grimbal meant.  Unfortunately, a truly open-ended education is probably never possible.  Even in my school, where students are allowed to make up their own majors, like "Rollercoasterology" or "Evil," they must still set down guidelines and benchmarks and submit them to a governing body.  Their ability to meet these benchmarks determines their grade.

Of course an open-ended educational system would be great.  It would be much more fun to experience topics at our own pace and allowed to apply them to whatever we chose without having anyone making us do it.  That's why places where that could happen are usually categorized as Utopian societies.
Title: Re: A proposal for a more creative CS education
Post by towr on Nov 6th, 2006, 3:48pm
Of course, given a few hours to think about it. I have my perfect project.
I'll make a program that retrieves lists of proxies from the internet, and uses those randomly to retrieve my website at random intervals. Thus inflating the page-count to rediculous heights for however long a period of time I desire.
Little effort, guaranteed A+
Title: Re: A proposal for a more creative CS education
Post by towr on Nov 6th, 2006, 3:50pm

on 11/06/06 at 15:45:57, Whiskey Tango Foxtrot wrote:
Even in my school, where students are allowed to make up their own majors, like "Rollercoasterology" or "Evil,"
That sounds cool, what school is that?
I might try to put together a "Mad Science" major, muahahaa.
Title: Re: A proposal for a more creative CS education
Post by amichail on Nov 6th, 2006, 3:50pm

on 11/06/06 at 15:40:55, towr wrote:
My suggestion woul dbe to first let the students find and define a 'problem' well, then evaluate whether that makes a worthwhile project. And perhaps find someone qualified to judge it and guide them.
Just letting them do whatever isn't doing them any favour (procrastination in the face of not having a clear goal is mighty tempting). And grading on anything other than the quality of their result is plain insulting.

I find it insulting that the vast majority of computer scientists do not even acknowledge the importance of open-ended creativity, at least in their teaching anyway.
Title: Re: A proposal for a more creative CS education
Post by amichail on Nov 6th, 2006, 3:53pm

on 11/06/06 at 15:48:06, towr wrote:
Of course, given a few hours to think about it. I have my perfect project.
I'll make a program that retrieves lists of proxies from the internet, and uses those randomly to retrieve my website at random intervals. Thus inflating the page-count to rediculous heights for however long a period of time I desire.
Little effort, guaranteed A+

Google works hard to catch cheating like this because people using adsense could do something similar to make more money.
Title: Re: A proposal for a more creative CS education
Post by towr on Nov 6th, 2006, 3:55pm

on 11/06/06 at 15:50:46, amichail wrote:
I find it insulting that the vast majority of computer scientists do not even acknowledge the importance of open-ended creativity, at least in their teaching anyway.
If they can decide and define their own problem, then it's open-ended, isn't it?
Or do you just want them thrown into the deep end without anything to clamp onto?

I don't see the sense in just letting them do whatever without helping them recognize when they're busy with a project that can't go anywhere. You're supposed to be teaching them, not letting them drown. They don't need to be at university of you're not going to be involved in what they do.
Title: Re: A proposal for a more creative CS education
Post by Whiskey Tango Foxtrot on Nov 6th, 2006, 4:14pm

on 11/06/06 at 15:50:45, towr wrote:
That sounds cool, what school is that?
I might try to put together a "Mad Science" major, muahahaa.

NYU, aka "The Ridiculous Major Machine."  I guarantee you could study "Mad Science" there and get your degree.
Title: Re: A proposal for a more creative CS education
Post by amichail on Nov 6th, 2006, 4:17pm

on 11/06/06 at 15:55:40, towr wrote:
If they can decide and define their own problem, then it's open-ended, isn't it?
Or do you just want them thrown into the deep end without anything to clamp onto?

I don't see the sense in just letting them do whatever without helping them recognize when they're busy with a project that can't go anywhere. You're supposed to be teaching them, not letting them drown. They don't need to be at university of you're not going to be involved in what they do.

This proposal concerns mostly the honours thesis project and its evaluation. Students could of course take standard sorts of CS courses as well. So it's not like they are completely on their own throughout the degree.


Powered by YaBB 1 Gold - SP 1.4!
Forum software copyright © 2000-2004 Yet another Bulletin Board