I assume you are in computer science, since you are talking about top-tier conferences. So, I am going to talk from a pure CS perspective, since in CS in order to get a PHD you need to have publications.
First of all, do not let your ego get in the way of you getting a PHD. For getting a CS PHD, you need a critical mass of good and solid publications. That means a sufficient number (not just one) of good to excellent papers. According to the CORE conference ranking:
Conferences are assigned to one of the following categories:
- A* - flagship conference, a leading venue in a discipline area
- A - excellent conference, and highly respected in a discipline area
- B - good conference, and well regarded in a discipline area
- C - other ranked conference venues that meet minimum standards
I know many people who took a CS PHD and went to successful careers afterward that during the duration of their PHD did not have any A* conference publication. Instead they had 1-2 A publications, a couple of Bs and some additional workshop / demo / poster publications. IMHO this is the most realistic plan for actually getting a good CS PHD and get it relatively fast. So, although it is good to aim high, you should know that getting an A* conference publication (VLDB, SIGMOD, FOCS, SODA) is not simply possible for all CS PHD students (and certainly not necessary) during their PHD duration.
In this sense, aiming for that one perfect publication that would take 4 years to write is counter-productive for you, because:
- If it gets rejected you start from null
- It is not enough to get you a PHD, because you cannot have a CS PHD thesis of 30 pages
- Everyone will assume that it is your supervisor's doing if you cannot follow through with equally good publications
- If you are Mr. Nobody in single-bind conferences, it will be harder for your work to be accepted initially, since nobody knows you or trusts you
- It is very difficult to do so, because writing a paper of this magnitude needs experience and this kind of experience can come only after writing many good papers and many rejection / resubmission circles for mere mortals like us (unless you are Terens Tao)
- If the paper is that good, it really does not make that difference where it was submitted. I know seminal papers presented in B-conferences with 300 - 1000 citations and papers in A* conference that nobody cites.
Also, your plan of sending to a journal is even worse, since it might take a year before it gets accepted and in the meantime many other people could catch up with you. And even if you did write this excellent paper, it will take some time for that publication to take off and be known and until then what? You will have finished a PHD with a citation number of less than 10. This is also bad on all counts.
On the other hand, putting out consistently good-solid publications in respectable A or B conferences on a frequency of 1-2 papers per year, means that in 3-5 years of your PHD, you may have 4-8 good publications and:
- People will know you for you (and not your supervisor) because you have
proven yourself consistently by producing good papers on regular intervals
- You will have gained better experience on how to actually write good papers and sell your ideas better
- You will have a better citation index, due to your citing your work
and other people citing it because the related community knows and respects you
- You will get review requests that will inform you faster on the current state-of-the-art
- You will find faster external collaborators which means even more good papers for you.
In a nutshell, my advice is: If your work has produced good-solid results, wrap it up and publish it, instead of aiming for perfection. Once your first work is out, everything else will be on its way.
