Productivity research material

Question

While debating which platform to use for what applications (specifically we debated c++, java, c# and f#) we tried finding scientific quality research on productivity (cost of ownership really) of similar applications build on various platforms/using the above mentioned languages. Any one know of such research?

EDIT: I see several have commented that the requested kind of research does not exist/cannot be executed. So I thought I'd clarified, since the kind of research I'm asking for is executable (and I know I've read one survey already but forgotten the source)

I'm not asking for a study that gives conclusion on what platform/language that would be the most productive "hammer" for any "nail". I'm asking for reasearch that given a set of precodition (as all research have) e.g. kind of HW platform, kind of application with comparable competencies in the staff, investigates on the tools (specifically language) and their impact on productivity. I'm not trying to start a religious war, that's why I'm not asking for peoples opinions but scientific quality research :)

Answer 1

I doubt you'll find conclusive research at that level.

Productivity and costs have a lot of factors including but not limited to your developers, the cost/ability of new developers on those platforms, the systems already in place, and other factors that are localized to your business.

It's not a problem for researchers its one for managers.

Answer 2

Been there, done that. This is the stuff religious wars are made of (in some places) ;-)

I doubt very much you will be able to come to a conclusive decision no matter how well you research the subject. All of the "platforms" you mention above have a lot of qualities and in the end, whichever direction you go, you'll have to compromise.

Don't exclude the option of using several different platforms depending on the circumstances (e.g. in one particular context I've regularly used C++ together with Java). There are many ways to skin a cat and interoperability is not by definition unmanageable.

Answer 3

Scientific research? I do not think so either. These would come from experience and FUD over the internet. Also, every vendor (IBM, M$,Oracle..) would have their own story of productivity based on their "reasearch". You may find something useful at PWC research site PWC innovation technology

Answer 4

I'm of the opinion that if it's purely productivity you're after, a team of application developers working in C++ will be less successful than a team working in C# or Java. Why: memory management, memory leaks, time consuming pointer errors, nasty bugs and crashes that have to be fixed, etc. I'm not even gonna say something about F# (is there a community of F# developers your programmers can turn to?). So, there, I've narrowed your selection down to either Java or C#. And that's as far as I'll go.

Answer 5

Productivity gains are dependent on application at hand. For a particular application there are clearly fits and misfits in using different programming languages or patterns. For example, using Object-oriented code instead of procedural code has clear productivity gains for a large software.Similarly, there are several patterns/anti-patterns for different applications. Scripting languages have clear programmer productivity benefit. I found this article which compares some of the programming languages for different tasks.

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.56.2977&rep=rep1&type=pdf

Answer 6

Not directly research literature but still interesting is the translation between function points and the number of code lines. This could give some idea behind the amount of work needed for the implementation of a single function point.

http://www.qsm.com/?q=resources/function-point-languages-table/index.html

Answer 7

Here's a C++ vs Java study.

Here's a semi-scientific study of Python/Django vs. C#.

And here's one on Java vs. C# vs. Ada.

Whatever languages you're interested in comparing, do a Google search for

study of [LanguageX] vs [LanguageY]

and see what you can glean.

But I agree with most everyone else that as long as you're in the right ballpark with your language selection, the rest of your productivity rests with the quality of your tools and people, not with whether you choose Perl or Python.

Answer 8

This isn't science, but it's based on 28 years in the business, observing and participating in teams which used a huge array of technologies, from assembly up through a long list of higher-level languages, plus of course a litany of methodological silver-bullet-du-jours:

What really affects productivity is your people - specifically, how good they are, and how well led they are.

So if you're trying to pick a language or other technology, go for the one with which more of your team is competent. If that's a tossup, or if it's truly green-field, hire the best people you can find, and then pick the tech best supported by the vendors whose tech you'll be dependent on... or toss a coin.

Answer 9

You have the answer to this question on your project already.

Take each of the languages you are interested in and ask each individual to rate themselves in that language from 1 to 5, with 1 being a novice and 5 being an expert in said language. Then tally up each language, whichever language has the highest level is the most productive language for your group.

If you are making a web app, strongly consider Ruby. If you are making a thick client, strongly consider C#. If you are just trying to make money consulting, use Java.