Does computer science research have value? (Inspired by the career path thread)

[GodlessAstronomer]

This thread went a little off topic and partially devolved into a discussion about the nature of computer science and whether research into the field has value in it's own right. Does theoretical computer science research for it's own sake have value in the same way that, say, pure theoretical physics has value?

My own opinion as a CompSci student in his final year of study is that CompSci isn't really a "science". I'll quote myself from the other thread:

Many believe that the theoretical advances should be made in the field, not in ivory towers. In my opinion the history of CompSci has shown this to be mostly true - the best research is done by companies or governments looking to make products, not for the sake of research itself. It's not like physics where developing theoretical models for the sake of knowing more about the universe around us has value in it's own right.

This doesn't include fields such as quantum computing, of course, but this is physics anyway.

What do you think?

 

[TruePaige]

I inspired a thread! 😀

To repost my response to kick start this:

I can definitely agree with you on that point, as I've talked with people who stand on both sides of the issue.

If I may play devils advocate for a moment though: How many people end up just toiling away on HIPPA compliance or creating the next Farmville / Mafia Wars clone though? When profit is the ultimate goal for most businesses research for the sake of research (especially in this day and age where significant advances such as artificial intelligence and embedded very small devices have become so complex) can we really expect non-research positions to be able to design these applications of the future?

Of course on the flipside: Practical applications do get some development traction such as the Hadoop project, and let us not forget the companies who have commercialized Linux, many other things such as the software to drive advanced 3D and holographic media has been advanced by private efforts.

So yeah, you can really see how it can be taken either way.

 

[apac]

Of course it has value. The internet wouldn't be here if it wasn't for research, and of course Al Gore.

My contention in the other thread was that computer science, as a science, is not limited to the purist opinion the other OP had. Computer science is largely an engineering discipline, rather than a science. All engineering has research.

 

[McCartney]

Of course it has value. The internet wouldn't be here if it wasn't for research, and of course Al Gore.

My contention in the other thread was that computer science, as a science, is not limited to the purist opinion the other OP had. Computer science is largely an engineering discipline, rather than a science. All engineering has research.

And nowhere did I dispute that Computing Science contains an area of applicability. There are many papers that USE what we have now in NEW ways and contribute to CS significantly.

One example would be using the Minimum Spanning Tree algorithm to develop a 2 approximation algorithm for the travelling salesman problem.

Or Edmonds Karp's shortest path heuristic for flow networks. Flow networks are pivotal to many areas of science. They developed a better heuristic that amoritzed the running time of Flow algorithms to O(EV^2) instead of, possibly exponential, when dealing with Ford Fulkersons.

 

[DaveSimmons]

From reading Communications of the Association for Computing Machinery the papers and overviews are a mix of interesting and somewhat pointless paper-generation.

Some of the interesting work is not likely to be done outside of a university since there is no immediate payoff, no instant practical application.

Some of the articles by industry practitioners also seem more shallow and less rigorous, not much above the level of "we did it and it's worked so far," without thinking too deeply about the theoretical side as a university researcher would (hopefully) do.

So, IMHO both "ivory tower" and industry research have their place.

 

[grohl]

need moar nerd research

 

[apac]

From reading Communications of the Association of Computing Machinery the papers and overviews are a mix of interesting and somewhat pointless paper-generation.

Some of the interesting work is not likely to be done outside of a university since there is no immediate payoff, no instant practical application.

Some of the articles by industry practitioners also seem more shallow and less rigorous, not much above the level of "we did it and it's worked so far," without thinking too deeply about the theoretical side as a university researcher would (hopefully) do.

So, IMHO both "ivory tower" and industry research have their place.

Agree.

 

[GodlessAstronomer]

Bumping this thread because I'm really interested in what you guys think about this

 

[chusteczka]

The only way this question could even be serious is if the OP has no conceptual understanding of the mathematical theory behind computer science.

Edit: mathematical theory

 

[FelixDeCat]

The science of computing?! How many ways can one explore a process to derive value? While the process can be improved upon, it remains essentially the same.

So I vote no.

 

[GodlessAstronomer]

The only way this question could even be serious is if the OP has no conceptual understanding of the mathematical theory behind computer science.

Edit: mathematical theory

Maybe the thread title was phrased clumsily (should be "how much value does pure comp sci research have") but you're dead wrong on your accusation. If you bothered to read the OP you'd know I'm nearly finished my CompSci degree and have taken more than my share of maths papers.

 

[chusteczka]

Maybe the thread title was phrased clumsily (should be "how much value does pure comp sci research have") but you're dead wrong on your accusation. If you bothered to read the OP you'd know I'm nearly finished my CompSci degree and have taken more than my share of maths papers.

I do not wish to be knocking you down but your attitude in the other thread was completely out of line and clearly displays your theoretical inexperience. Your discussion belittles your own "nearly finished CompSci degree" and shows your school does not provide enough for a solid, in-depth theoretical computer science experience.

How do you think the Oracle optimization engine gets developed? Oracle gathers all the leading published theoretical research, adds to it, and then implements it.

Then how about all the research that leads into algorithmic theories regarding the traveling salesman problem, automata theory, and graph theory? TomTom GPS would be lost without such theoretical research. Are you saying Knuth and Dykstra are irrelevant to life? Your question and attitude drops my jaw to the floor and leaves me drooling like a Saint Bernard.

 

[RESmonkey]

i think you would be more happy making bank doing software engineering then being a professor at a university doing papers and research (at least I would be happier, even more if it were game dev)

 

[McCartney]

I do not wish to be knocking you down but your attitude in the other thread was completely out of line and clearly displays your theoretical inexperience. Your discussion belittles your own "nearly finished CompSci degree" and shows your school does not provide enough for a solid, in-depth theoretical computer science experience.

How do you think the Oracle optimization engine gets developed? Oracle gathers all the leading published theoretical research, adds to it, and then implements it.

Then how about all the research that leads into algorithmic theories regarding the traveling salesman problem, automata theory, and graph theory? TomTom GPS would be lost without such theoretical research. Are you saying Knuth and Dykstra are irrelevant to life? Your question and attitude drops my jaw to the floor and leaves me drooling like a Saint Bernard.

God bless your soul. And this is why I'm trying to show restraint and impartiality because of the fact that many of the up and coming CSers don't share the same attitude as you and I, which is fine. I'm not saying he should be doing research until he's older, or he needs to contribute. All I'm saying is that our fundamental purpose is to understand the theory and mathematical theory to be efficient so that we can OPTIMIZE many pieces of software.

There is a science here. But thank you so much for at least sharing the same rough idea.

 

[gimmewhitecastles]

All sciences have some sort of profit motives attached to it for certain people. Even the natural sciences (physics, chemistry, biology) can be applied to military applications that tend to bring in a lot of money. There may be some ethical dilemmas that may have to be squashed in order to proceed in this field but sometimes the power of the fat paycheck will eventually win over.

And anything can be called a "science" if there is more knowledge to gained and researched. The definition of science speaks for itself.
http://www.merriam-webster.com/dictionary/science

So yes. Computer Science is in a valid science.

 

[ivan2]

more like arguing if math is a science or not.

 

[chusteczka]

Public research advances the state of the art for everyone's use. This includes theoretical computer science research.

The government often pays for aspects of technology to be developed in areas that a private company will not be interested in due to a low Return On Investment (ROI).

As an example, look at all the math behind the calculation for ROI in the above link.

 

[Legendary]

I do not wish to be knocking you down but your attitude in the other thread was completely out of line and clearly displays your theoretical inexperience. Your discussion belittles your own "nearly finished CompSci degree" and shows your school does not provide enough for a solid, in-depth theoretical computer science experience.

How do you think the Oracle optimization engine gets developed? Oracle gathers all the leading published theoretical research, adds to it, and then implements it.

Then how about all the research that leads into algorithmic theories regarding the traveling salesman problem, automata theory, and graph theory? TomTom GPS would be lost without such theoretical research. Are you saying Knuth and Dykstra are irrelevant to life? Your question and attitude drops my jaw to the floor and leaves me drooling like a Saint Bernard.

To use your own example, my main problem with the OP from the original thread is that he belittled those who converted graph theory etc. into a TomTom (people who program for a living)

No matter how great a theoretical idea, researchers still need people with practical knowledge to apply those concepts.

As a DBA I would rather deploy the first instance of a new database engine into a true production environment and really put it through the rigors of reality. Developing it via reading papers and 'lab' research does not appeal to me as much. You and McCartney might prefer the opposite.

Now if you think that makes you better than me then we must be enemies 😀

 

[chusteczka]

...
My contention in the other thread was that computer science, as a science, is not limited to the purist opinion the other OP had. Computer science is largely an engineering discipline, rather than a science. All engineering has research.

"Computer Science" is an engineering discipline strongly based on mathematical theory and scientific research. The results of this theory and research is included in the software libraries that comprise the programming environment that software engineers, programmers, and information technology technicians access to accomplish their tasks.

 

[Kirby]

research has no value unless it can be used. theory and research bore the hell out of me, i like getting shit done. make something happen.

 

[McCartney]

To use your own example, my main problem with the OP from the original thread is that he belittled those who converted graph theory etc. into a TomTom (people who program for a living)

No matter how great a theoretical idea, researchers still need people with practical knowledge to apply those concepts.

As a DBA I would rather deploy the first instance of a new database engine into a true production environment and really put it through the rigors of reality. Developing it via reading papers and 'lab' research does not appeal to me as much. You and McCartney might prefer the opposite.

Now if you think that makes you better than me then we must be enemies 😀

Graph theory doesn't interest me. I didn't belittle anyone who did it. I have quite a few friends doing it. It's just very, very different and requires incredible skills. I love Algorithms and Complexity theory just not like the people doing Phds or MSc's in it.

 

[chusteczka]

To use your own example, my main problem with the OP from the original thread is that he belittled those who converted graph theory etc. into a TomTom (people who program for a living)

No matter how great a theoretical idea, researchers still need people with practical knowledge to apply those concepts.

As a DBA I would rather deploy the first instance of a new database engine into a true production environment and really put it through the rigors of reality. Developing it via reading papers and 'lab' research does not appeal to me as much. You and McCartney might prefer the opposite.

Now if you think that makes you better than me then we must be enemies 😀

I completely agree with you. This topic is funny for me to be defending since I am not that strong in theory. There are many others far stronger than I am. I just went to a good university that taught theory rather than practical application. I have the experience of being taught in such an environment. I had trouble convincing employers that I could program since I had a substantial theoretical understanding but not as much practical understanding as other job applicants from other schools.

You seem to have picked up on a bit of inexperience on the part of the OP from the other thread. This is why we discuss such concepts here, a desire to learn more about the shady areas of our own understanding. To round out and enhance our personal knowledge. This includes the OP from the other thread and the OP from this thread.

From my experience, there were researchers isolated in their own little world. These researchers were not well liked under the surface and they made poor instructors. There is no way I could call them a "teacher". They truly showed the unpleasant distinction between "teacher" and "professor".

Then there were the professors that were involved with private companies or a group large enough to provide a strong practical focus for the research and development of strong programming libraries. These professors were excellent and provided strong tools for others to work with. Off the top of my head from my limited experience, such names include Gropp, Heath, Kale, Kamin, and Ralph Johnson a Gang Of Four member that developed the concept of design patterns. These are all strongly theoretical researchers who have had a strong practical impact on scientific development. As an example, Professor Johnson had strong collaboration with numerous researchers and private, application focused contacts when he helped develop the design patterns that have permeated current software development practices.

Most theoretical research is funded or otherwise focused to solve a practical problem.

This discussion shows me the worth of having received 'C's at a strong theoretical university rather than 'A's at a more practically focused university. The wide understanding and theoretical experience I received allows me to see outside of the practical boundaries inherently imposed by solving practical problems within a pre-built software environment provided by the theoretical researchers (i.e. Java, C++, Oracle).

 

[Red Squirrel]

Nothing beats hands on, on the job experience. College/university is just to get the papers required to even apply for a job, it does not help much on the job. Every environment is different and will have stuff to learn.

 

[chusteczka]

Nothing beats hands on, on the job experience. College/university is just to get the papers required to even apply for a job, it does not help much on the job. Every environment is different and will have stuff to learn.

God, this is frustrating! What part of the educational process did you miss to think that a university education does not help much on the job? Specifically, you often ask largely scoped questions in other forum categories in an attempt to solve problems that have already been solved. You do not have the theoretical experience to understand its use.

Why are those without the theoretical experience so determined to put theory down when your questions and discussion itself prove your incorrectness on this topic?

 

[chusteczka]

There is a reason the 'A' students go into theory, the 'B' students go into solving applications, and the 'C' students look for other ways to earn a living.

Theoreticians anticipate practical problems and create tools to solve these problems.

A well-balanced mix between theory and practical application is necessary. Insufficient understanding occurs when this mix gets out of balance.

I was under the impression the non-theory schools teach enough of the theory for a strong practical understanding of all areas. Maybe this balance between theory and practical skills displays the difference between engineering universities and IT trade schools, since I have no idea of the educational levels of those catching my ire at the moment.