I need to change PhD topics. Should I pursue one of my advisor's ideas, or one of my own?

Question

I had an idea for a PhD proposal and found a supervisor for that topic. My supervisor is pretty busy and is reachable only once every month or two at best.

Unfortunately, I saw that my research topic could be usable only for publishing a paper, but not to develop something really novel that could justify its presence in a PhD thesis. Thus, I feel like I've "wasted" a couple years of my PhD.

When I realized that my current research was not enough for a PhD thesis, I came up with another idea related to the same field and topic, but the problem is that my supervisor is not an expert in the field that I am proposing. To discover this other technique I spent an additional year.

Now I have the following courses of action:

• Since my supervisor is not an expert into the field that I am proposing, then I should try to find a co-supervisor at another university. This is not so easy because a lot of professors have their own set of students to advise.
• My supervisor has a list of projects that could serve ås PhD research topics, but I'm not so interested in those topics. However, I need to graduate so I am seriously thinking of picking one of them. The problem is that I am very rusty in the background of the topics he has because since I haven't looked at those topics since my bachelor's program.
• I am pretty confident that I can continue to develop my new idea by myself, even though I know that is going to be pretty difficult. The problem is that I will have to do it all by myself, and my supervisor will not be able to give me any advice because it is not his field of expertise.

If anybody is curious my PhD is in Computer Science.

So what should I do? Should I aim for a new topic or should I stick to my "feelings" that I can make my proposal work, even I have to do it alone.

Answer 1

Big Fat Note On Timing: First and foremost here, make sure to changing your topic is allowed by your supervisor. Then your supervisor will check the time you have left and will point out if you can do it or not. You don't want to change a topic while your supervisor do not know about it.

Answer On Changing the PhD Topic With Your Supervisor Agreement

Ok, I did my PhD in CS as well; also was very close to change it for two times as my supervisor was not as an expert I thought he would be. So here my takes on changing your topic as a CS PhD student:

1.What you have gained so far?: Without comparing yourself to anyone else, are you deep down happy with your progress and how much did you gained, knowledge wise, so far? The conclusion in this point brings me to the second point.

2. Problem is Environment or Yourself?: After thinking about how much you have done, if the sadness relates to the environment (such as supervisor) or yourself (such as the issue of focus).

2.1 If the problem is you: If the problem is you then, changing your topic does not change a thing. Now you might say well "I will change myself"; this is not as easy you might think and you might need professional help on this. If you see you need help and the supervisor is not helping either or even recognizing that you are having an issue, then you need to think seriously on how to get help. This is very important, as you might get depressed without doing anything wrong. You see what I'm trying to say here?

2.2 Problem Is Supervisor: It is more common as you might think, so in this case, if you have the ability to do the PhD, then YOU need to inform the supervisor about the changes and YOU need to decide what you want to do. This might get ugly at times, because if the supervisor does not know the topic and tells you, your thesis is total rubbish, you should know "well I'm submitting it and defending it regardless". So basically, take the control of the situation from start to finish.

2.3 Problem is Topic: To be honest I doubt it, because most of the time the problem is based on the first two sub points. Anyway if this is the case, then change the topic. Don't listen to supervisor to do a topic in a field he/she did not publish a thing for the last 5 years. Be honest to yourself and your supervisor ability before choosing it. This should be also discussed with your supervisor as well, don't accept anything without knowledge. I recommend to choose a topic that your supervisor is known for. If your supervisor works on software testing and you want to do a topic on code generation, things might not go well.