Should I aim at publishing a single-author paper on my PhD?

Question

This is the situation. I'm a 2nd-year PhD student in Computer Science. The research group I belong to is very large, and during my first two years I have worked with many different people from the group, and also gathered a lot of contacts from other universities. In fact, I haven't really met my supervisor, let's name her Alice, for anything besides administrative stuff.

Now, I have been working on something with professor Bob, who works in a different university. Most of the work has been done by me, and since Bob is now out of time he said it was better for me if I continue this work with Alice. The thing is that I've been doing this pretty much on my own from the beginning, and I know I could finish it on my own as well. Also, as I said initially, my group is very relaxed in terms of student-supervisor relation, so I'm thinking of just continuing the project on my own and submit a single-author paper to X conference.

My specific questions, given the above, are the following:

• Should I pursue this on my own and just do it without notifying my supervisor? (it's important to say that in my group this happens often, but not for single-authored papers; i.e. a lot of students work with many other professors somewhere else without notifying their supervisors, and that's all fine).
• If you think I should talk with my supervisor instead, then how should I approach her? "Do you want to get involved in this" or "Can I work on this on my own"? (maybe it's relevant to say that my supervisor has co-authored some papers with me in which she hasn't really done much, not even the first-core idea, so maybe the same principle should apply here?)

I know this is very much dependent on how things work at my place, but I would love to hear the advice from other researchers as well.

Answer 1

All of the doctoral work I'm most familiar with in CS has been published as single author work. You may want an extensive acknowledgements section if you have worked with others. There may be some corners of CS in which there is some division of labor, such as complicated data gathering v data analysis algorithmics that would be an exception, but I think that is still rare.

Even in cases in which an advisor supplies the initial problem, it is fine in most of CS to publish as a single author. Mathematics is similar. It is in the lab sciences primarily that the standards are different as the very existence of the lab is essential as is collaboration with other lab people.

But yes, speak to your supervisor and lay out your reasoning as here. You might even write up a draft that includes the ack section if you are far enough along. What you want to achieve, however, is a plan that makes everyone happy (enough) and leaves your advisor as a supporter of your work and future.

But if you are in one of those subfields in which such multiple authorship is the norm, then your advisor should be able to convince you of that.

But a rule of thumb might be the distinction between "I built this and you helped" vs "We built this together". There is a fuzzy area between them, of course. In general, though, it is best to know in advance which situation you will fall in to.