Let's do more computer science | Hacker News Reader

Let's do more computer science

scripting.com

69 pointsdavewiner24 comments

I'm a sophomore in college and I agree. Emphasis on making mindless web apps has made my peers (and me, sadly) into consumers of SDKs and frameworks, rather than producers of technology.

To add to the confusion, every VC with a blog is writing about how CS education is irrelevant and that it's worth dropping out. But I doubt that the core tech behind Google or the erstwhile Sun could have been made if their founders hadn't learned proper computer science.

Google is not like the current wave of VC web apps: it was built on an idea that couldn't exist without the proper technology. For most VCs, that technology is merely "the internet," and their ideas simply leverage web connectivity to get up and running so they can see how they do in the marketplace. If a stronger technological backbone is needed later, they can always hire computer science grads.

For example, Facebook didn't need computer scientists to start out, but it needed them to grow. For VCs looking to make a buck, the business idea comes first, and innovation in technology can follow if it needs to.

One comment I would have on this is that theory is often not framed by many professors in the context of practice or application. The converse is also true.

Math theory is a beautiful thing, but if no context is provided (this can be merely history, or a story of how it was applied, rather than actual applied homework itself) would go a long way with students. Otherwise you're just teaching people the notation and rote plugging and chugging of theory, rather than true understanding of the theory.

As a counterpoint, I think the appeal of (mindless\thoughtful) web applications are keeping people interested in Linux, open source software, new languages, and motivating them to continue learning on their own.

The vast majority of non-top tier CS programs just manufacture Eclipse\Visual Studio users for large companies without motivating students to keep discovering new skills. Focusing on a couple successful outliers isn't a representative sample.

Selection bias much? He's complaining about an over abundance of consumer facing web businesses at an event sponsored by a blog that focuses almost exclusively on consumer facing web businesses.

There is lots of real computer science going on in the world, and there's a bunch of money being thrown at it. It's just that you won't hear about any of it by reading TechCrunch all day.

On which sites will I hear of this "real computer science"? I'd be interested in reading them along with my daily dose of HN and TC

It's not going to show up popular websites much because CS research, like most research, is not a popular topic. But if you want to know what's going on, a great place to start is to start reading the proceedings of recent ACM conferences: http://portal.acm.org/browse_dl.cfm?linked=1&part=series...

Yes, it's behind a pay-wall. Not much I can do about that [1]. And yes, you will have to read academic papers. But that's where people talk about real computer science.

[1] Although, if you find a title and abstract that sounds interesting, you can usually just Google for the title and find a copy on the author's site. All of my papers, for instance, are also on my webpage. And, yes, this is explicitly allowed by the copyright transfer agreement.

In some specific areas (like mine: graphics and vision), there are compiled lists of recent papers from the best conferences, along with links to the papers, project pages, etc.:

http://kesen.realtimerendering.com/sig2010.html

http://www.cvpapers.com/cvpr2010.html

http://www.cvpapers.com/eccv2010.html

http://www.cvpapers.com/iccv2009.html

I have been a member of ACM for the last 2 years, but just let my membership lapse. Their site is nearly useless for anyone who is not already at least an advanced undergraduate in CS or IT. The parent post is looking for resources for learning CS, not advanced study (or so I understood it).

http://lambda-the-ultimate.org/ is worth a look!

When you find a really good site on HN, add it to your RSS reader. Check out proggit. If you have an industry you're interested in or a problem you're trying to solve, start googling and keep track of all the interesting sites you come across. There's a wealth of good stuff out there, though it isn't going to be TC style (linkbait titles, unnecessarily effusive prose, etc).

Uhh what does TechCrunch have to do with this?

Did you read past the first paragraph of the piece?

Come on fess up, you didn't did you? ;-)

We used to do that, when I was a comp sci grad student in the 1970s.

And that still happens... in grad school.

I am confused by this essay - I'm not sure what his main point is. I'm going to guess it's "we're not doing enough computer science." But his example for this being the case is a conference focused on startups, which is a complete non-sequitur to me. It's like sitting in a biology lecture and wondering why these people aren't doing more physics.

We're doing a hackathon at NYU and if it's like last year's it will be replay of the kind of hackathon they do at TechCrunch. I'm saying our students should be encouraged to do projects that aren't just subsets of whatever was hot with the VCs and angel investors 8 months ago. I thought I was pretty clear about that in the piece, but then you'd have to actually read it to figure that out.

I read the piece - twice, actually, because I was looking for a clear thesis. I still find it strange that you want students to do as you did in grad school at an event focused on startups. If you want to find oranges, don't look in an apple orchard.

I want them to use the computer science skills they learned at the university that is holding the hackathon. Is that simple enough for you?

> We used to do that, when I was a comp sci grad student in the 1970s. I think we got way too caught up in the commercialism.

Students choosing to study computer science today are motivated by different reasons than their counterparts from the 70s. Many choose to study computer science at the undergraduate level specifically for these commercial aspects that were not present 30-40 years ago.

There is a good reason why good computer science doesn't come out of a hackathon type setting. Good computer science that shows us something new really requires at least 6 months of solid work to get up to date on some of the research in a narrow field. Even after over 6 months of reading academic papers I realized I would have to keep my focus narrow because some of the other things I thought about would have required much more reading.

In a short period of time you can't really hope to produce anything meaningful in new ground, you can however produce interesting little ideas and mashups which is what hackathon like events tend to focus on.

I agree with you wholeheartedly, and I would hybridize what you said a bit.

There are always amazing things being studied in computer science, and those are a huge opportunity of things to be monetized.

Just as things like the web browser and pagerank came out of university research, we should be looking in academia for the next massive opportunities in technology.

The author points that programming projects are driven by or looking for venture capital, so there are few computer science projects.

What is the solution? You can't pay creative people to build computer science. The only think that society can do is to increase the respect and consideration for those that strive to widen the frontiers of knowledge. The great work will be done when those men reflect a flame in their eyes.

Inspiration is the force of gravity for discovery, but not everyone lives in the same planet here. Allow people to grow and make research and you should find inspiration and the force to discover and innovate.

I'm not a coder/hacker/tech person and largely ignorant of the issues that are being discussed in this post. But, I'm curious: is some of this similar to the (mostly light-hearted) tension I've seen in my personal life between physicists and engineers?

I don't think it's meant in that way. I think the author fears that young computer scientists are being coddled - instead of encouraging them to learn how to solve hard problems, they are rewarded for producing work with little substance but lots of polish.

The main issue with this is the people that are complaining about the up coming generation of computer scientist are doing nothing to improve the situation.

Why isn't there a site dedicated to computer science like the khan academy?

Why aren't you creating resources for students that can help us understand why the fundamental concepts of computer science matter?

We obviously need help.

I'm not a computer scientist, but I would like to point you to the MIT OpenCourse site (http://ocw.mit.edu/). Or any number of similar sites. Or even algorithm competitions like ACM ICPC/Topcoder.

Reply on news.ycombinator.com