Author refuses sharing his "open access" code/data

Question

My doctoral project required me to learn a whole new method from zero. My advisor had another student who mastered the technique since his BA studies, and so was much more advanced than me. I asked this student several times to share his code with me, but he claimed that his research was "classified" (No, it was not), so he is not allowed to share it. Now his papers are published in open access journals under the statement

Scripts using the R and MATLAB programming language are provided to produce figures and extract data from external databases. Additional code and related files are available from the corresponding author (him) upon request.

Only that now he is ignoring my emails.

Is there anything I can do? The truth is I want to use his codes simply because I want to learn from him.

Answer 1

Here are the typical steps to follow until you got the code:

1. Email the student who has the code. Send a reminder after a few days.
2. Email the student who has the code and CC the student's advisor and/or co-authors. Send a reminder after a few days.
3. If you can meet either student or better the student's advisor face-to-face, try it.
4. If none of the co-authors is your advisor or close to you, email the journal to raise a formal complaint that the author didn't send you the code. Attach your previous emails to the student who has the code as a proof. If one of the co-authors is your advisor or close to you, explain the situation and get them to put pressure on the student who has the code, to comply with the open access requirement of the journal.

That should resolve the issue in most cases.

---

PS:

His papers are published in open access journals.

Note that "open access journal" doesn't entail open source code. Neither the journal nor the student who has the code has any legal obligation to share their code (unless some special cases, e.g. if the student's funding agency forces them to do so in their funding clauses).

Answer 2

I interpret the question as "I need code to do stuff, how?".

• The classic option, the one that you are doomed to use if everything fails, and the way things were before open code repositories, is: just hire / take a BSc student. They would need to implement that thing from scratch using papers / books on the topic as a part of their job / BSc thesis.
• Ask the PI on the project / the corresponding author of the paper, if they would share the code with you. They might not agree. I am sure, other answers will detail on the open access and availability bit, I suspect some internal shenanigans.

Basically asking for code is not the best way to do it in my opinion. You do not want to study the code, you want to do stuff, right? Then ask about it. There are multiple options.

• "I need your code, because I want to understand how it works / reproduce the results in the paper." That's what the availability thing is for, officially.
• "I want to build upon your code to produce new research results." This is the most favourable version. Mostly implied with it (but it does not hurt to state it explicitly) is: "you get to be a coauthor on the new paper".
• "I want to apply your method to my data and to use the results in my paper." It is a milder version of the above. Some authors are not inclined / not allowed to give the code to anyone. (This behaviour becomes less frequent, for good, but is still present.) Still, they might be interested in a collaboration. A possibility is to give them your data (if you, on your end, are comfortable with it), let them process it, obtain the results. Again, it implies "you get to be a coauthor", as they do work for your new paper.
• "I want to compare your method to mine." It is a valid concern, in most of the cases it produces just a citation of the others' paper.
• "I want to do some statistics / automatic code analysis on your code." That's a citation, I'd say.

I'd say that people are in generic more inclined to provide data or to collaborate on one of the above terms that are more exact and clear than just "let me see your code".

I would assume that the most frequent "fears" that lead to not giving code to other people are: "they might 'steal' it" – i.e., an idea that someone would build upon your code and not acknowledge you; and "legal" limitations, ranging from "I have transferred the copyright to my research unit and they don't share" to "my supervisor said 'no'". (Again, I am not discussing here, how the data availability statement of the publication agrees with those.)