17

I am a PhD student in Management, at Autonomous University of Barcelona.

While working on my thesis, I found that there is no piece of code or software related to a specific but important topic of efficiency analysis and benchmarking. I am really eager to invest part of my time and write a package in R in order to help other researchers as well as deeply learning the topic through coding. Moreover, when I publish my paper I can motivate interested researchers to use that piece of code and re-evaluate my work or extend it.

Now my question is whether writing such package equivalent to a chapter of thesis /a scientific paper. Based on the answer, I will decide whether to work on it right now or to defer doing so.

I will be grateful for any help you can provide.

  • 15
    This is something you should really ask your advisor about. Would your code contain novel algorithms and techniques, or would it just be an implementation of something tried and tested? In the first case you would likely end up with something publishable, while in the second case you wouldn't — although it would probably be a useful learning experience.. – A.P. Dec 27 '15 at 10:25
  • 5
    You could always make the package as opensource software and then publish it as a software journal article. See my answer here. Full disclosure: I am associate editor with the journal. – dgraziotin Dec 27 '15 at 10:41
18

Writing useful software is not equivalent to writing and article or a thesis chapter (which are generally not equivalent to one another either). It may, however, be equally valuable for your career, depending on what you want in your career.

The real question is whether people will find it highly valuable. If they will find it valuable, then it can be worth something to you. If you want it to readily count towards a conventional academic career, then after the software is done you can write a journal article about it, and people can cite that. If you search for "analysis software" in Google Scholar, you'll find many highly cited works of this sort, showing that it is certainly possible, even if not the straightest and most conventional road to take.

  • 3
    I would also suggest putting something on the website asking people who want to cite the software to cite the paper you wrote about it. – Peter Green Dec 27 '15 at 15:00
  • @PeterGreen Like GNU parallel does, for example? – muru Dec 27 '15 at 20:20
  • +1. Of course, the logical place to publish that journal article accompanying the package would be the R Journal. – Stephan Kolassa Jan 4 '16 at 14:49
3

Writing an R package is not necessarily the equivalent of a PhD thesis chapter or a journal paper. The best to answer those questions would be your current PhD advisors. However, for your own career it all depends on what will be the respective quality, impact, or influence of your R package vs. your completed PhD thesis. A successful R package does give one incredible credibility within the quantitative fields, sometimes even as much or more so than a PhD. Let's put it this way if two PhDs are pretty much equal in every respect but one has demonstrated a superior knowledge and impact on the R community with an innovative and successful R package related to his field of expertise; the PhD with the R package will be considered a lot more valuable than the one without. Given that, I don't think your question is either do this or do that. I think you should do both. Complete your work to earn your thesis in the most traditional way. But, also take the extra effort to develop that R package in a manner most relevant to your field with also general application. This is going to take you more time. But, it is so worth it. If you don't do it now when you have the liberty of contemplating budgeting the time for it, you may regret it later when professional time pressures may prevent you to ever again reopen this opportunity.

3

As a PhD advisor I would count R package as a journal paper. At least if I would have to make an evaluation of your thesis :-).

The answer, however, depends on your university. Why don't you write a package an publish it in a journal? Using it as part of your thesis is of course permitted. And if you get other users on board, everyone can see the importance of your work.

Although this is not established as a standard, I would encourage you to do so.

  • Many many thanks KrOstir. I am really enthusiast to do so because without a publicly available code, very few researchers are able to work on the topic. Thanks again – Shahin Dec 28 '15 at 12:32
1

Writing software is not equivalent to writing a thesis, because the thesis is a theory defended by a practical demonstration.

It could be equivalent to the research project about which a thesis is written. Any novel elements could be the parts presented as the question to prove. (algorithms, architecture, or even methodology)

(Your committee may actually allow you to pretend otherwise, but that really is more of a non-thesis option if you never actually write a thesis and defend it.)

The classical approach to this question is to write up and present a research proposal to your panel. If they accept it, it becomes your official thesis topic. Obviously, as already suggested, most bounce the ideas off of their adviser and peers before the formal proposal.

  • Thanks a lot "The Nate". Indeed, writing a package was neither part of my proposal nor my crude ideas when I started to work on the topic. Then in the last summer I figured there is no publicly available piece of software for that specific purpose. My idea is not substitution of the thesis with the R package, but presenting the package as just a chapter of my work. So two chapters are theoretical and methodological contributions, then one chapter is the software in order to make those contributions publicly testable, available and more visible. – Shahin Dec 28 '15 at 12:37
  • That sounds likely to pass your committee. Tweaking research after proposing is normal. – The Nate Jan 4 '16 at 13:53

protected by ff524 Dec 27 '15 at 19:10

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).

Would you like to answer one of these unanswered questions instead?

Not the answer you're looking for? Browse other questions tagged or ask your own question.