When I read the original poster's story, I felt this smells of poor supervision.
A sympathetic supervisor might be a compassionate/"nice" person, but still a poor manager: he should have immediately objected to set up a monster team of students; it's hard to succeed in teams of professionals, but students' goals are not aligned with your own, so they bound to walk away eventually. Nobody can blame them for prioritizing their exams. That's why there are Research Programmers.
What I conjecture might have gone wrong is nobody helped the OP to divide the
original project idea into manageable chunks of work that lead to sub-results that lead to smallest publishable units. A Ph.D. thesis is a marathon, not a sprint, and you need to plan accordingly, with milestones every now and then so that the supervisor (and his funders!) can be fed interim results in the form of publications. These milestones also motivate yourself, and they de-risk your project, so that at the end, your thesis can be produced by (slightly simplifying here but only slightly):
Think about what sub-questions your overall research question can be split into?
What code do you need to write and what experiments do you need to run to support your "central thesis" (which one poster in the above writing advice rightly recommended should be formulated as a question with sub-questions)?
In the papers I read about procastrination, the reason commonly given is being overwhelmed by a huge monolithic problem. The good news, then, is: you as a computer science know "divide and conquer" the technique that is the cure for monolithic disease - create little chunks that you can solve, publish, and use as building blocks. Each of them won't be scary. It follows that a finite sequence of non-scary steps won't be scary.
Ask experienced researchers around you for help in dividing up your research into manageable, implementable, publishable chunks. Remember: A Ph.D. thesis is an argument that your "central thesis" holds, so think hard what you need to show and how you can decompose bigger items into sub-questions. I wish you success, and most of all, stay sane!
I know you can do it, because I read your very well articulated post, and problem analysis is 90% of the solution, you also seem to know how to ask for help when you get stuck.
When I read the original poster's story, I felt this smells of poor supervision. A sympathetic supervisor might be a compassionate/"nice" person, but still a poor manager: he should have immediately objected to set up a monster team of students; it's hard to succeed in teams of professionals, but students' goals are not aligned with your own, so they bound to walk away eventually. Nobody can blame them for prioritizing their exams. That's why there are Research Programmers.
What I conjecture might have gone wrong is nobody helped the OP to divide the original project idea into manageable chunks of work that lead to sub-results that lead to smallest publishable units. A Ph.D. thesis is a marathon, not a sprint, and you need to plan accordingly, with milestones every now and then so that the supervisor (and his funders!) can be fed interim results in the form of publications. These milestones also motivate yourself, and they de-risk your project, so that at the end, your thesis can be produced by (slightly simplifying here but only slightly):
Think about what sub-questions your overall research question can be split into? What code do you need to write and what experiments do you need to run to support your "central thesis" (which one poster in the above writing advice rightly recommended should be formulated as a question with sub-questions)? In the papers I read about procastrination, the reason commonly given is being overwhelmed by a huge monolithic problem. The good news, then, is: you as a computer science know "divide and conquer" the technique that is the cure for monolithic disease - create little chunks that you can solve, publish, and use as building blocks. Each of them won't be scary. It follows that a finite sequence of non-scary steps won't be scary.Ask experienced researchers around you for help in dividing up your research into manageable, implementable, publishable chunks. Remember: A Ph.D. thesis is an argument that your "central thesis" holds, so think hard what you need to show and how you can decompose bigger items into sub-questions. I wish you success, and most of all, stay sane!
I know you can do it, because I read your very well articulated post, and problem analysis is 90% of the solution, you also seem to know how to ask for help when you get stuck.