Software can be a primary product of scientific research

Once again, I’m writing in response to Titus Brown, who answered Is software a primary product of science? with “no” in his blog and on Twitter:

Note that Titus cleverly changed the question from asking about research to asking about science. Perhaps we can combine these and think about scientific research.

So, what are primary products of scientific research? I think most would agree that books and journal papers qualify, potentially along with book chapters, and conference papers, depending on science domain. So, writing text seems to be accepted as a primary product of research.

But this isn’t quite right: it’s not any text that counts; the subject of the text also matters. Generally, the text should be about a hypothesis and proof about the truth or falsity of the hypothesis, though it can sometimes also be a survey of previous work (adding new insight) or a position statement.

So what is the product? Is it the text, or the underlying knowledge gained? I think it’s the knowledge, as represented in the text.

How else can we represent knowledge?  In data? In software? In tools or instruments? The answer is maybe. Much like text, all data, all software, all tools, all instruments, etc. are not products of scientific research. But some are.

It may be worth considering the (translated) words of Alfred Nobel who left a large part of his estate to be given as, “prizes to those who, during the preceding year, shall have conferred the greatest benefit on mankind.” Reading further, the language that describes the criteria for the prizes in different fields varies, as follows:

  • physics: “most important discovery or invention”
  • chemistry: “most important discovery or improvement”
  • physiology or medicine: “most important discovery”
  • literature: “most outstanding work”
  • peace: “most or best work”

So, could a Nobel Prize be given based on software? It appears that the answer would be yes in physics or chemistry, but no in medicine. (This is assuming that the software is not a discovery itself, which might be the case in computer science, but that’s not a field Nobel knew about.)

If it’s worthy of a Nobel Prize, does that make it a primary product? Probably yes.

Is all software potentially worthy of a Nobel Prize? Certainly not.

Is all software a primary product of scientific research? No. But a lot of it, particularly software that is intended to be used by others, is. This type of software encapsulates knowledge and methods, and implicitly demonstrates the proofs of hypotheses.

Is this a clearcut distinction? No.

I admit that I mostly agree with Titus’s post about software’s importance and role in scientific research, including “better software, better research!” but I’ve weighed the same evidence and come to a different conclusion.

Is software a primary product of science? Overall, I think there is enough software that is a primary product of scientific research that I can answer yes.


Some work by the author was supported by the National Science Foundation (NSF) while working at the Foundation; any opinion, finding, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the NSF.


Published by:


Assistant Director for Scientific Software and Applications at NCSA, Research Associate Professor in CS, ECE, and the iSchool at the University of Illinois Urbana-Champaign; works on systems and tools (aka cyberinfrastructure) and policy related to computational and data-enabled research, primarily in science and engineering

Tags , 3 Comments

3 thoughts on “Software can be a primary product of scientific research”

  1. Well, I certainly wasn’t trying to be clever with the title – I was just clumsy (and it probably reveals some deep seated implicit bias, too).

    Thanks for engaging with me; I’m trying to understand your blog post still ;). Are you saying that we should think of software as a communication of methods and hypotheses to others? I could agree with that, I think. Do you think that software can be sufficient as an argument on its own, or does it need to come with other artifacts? And if so, any thoughts on what, specifically?


    1. Engaging with you forces me to write down my thoughts and makes me think a bit more deeply than I otherwise would 🙂

      I did assume the title change between the blog and the tweet was not a conscious choice, but following that thread was a different direction than I chose for this post. A discussion of how science research compares with other research would be interesting to have another time.

      I think that some software is indeed a communication of methods and hypotheses to others. And I think that good software (meaning a package that include documentation, tutorials, examples, etc.) of this type can stand on its own. Not that all software is for this (your previous Linux countertweet, for instance), and not that all software that communicates methods can do it without a paper.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s