The following answer is based on my experience during the last four years as a doctoral researcher in the area of distributed systems. I don’t know about practices in all different areas in computer science. Computer science is a big field and there are extremely large differences in the quality and quantity requirements of your Ph.D. supervisor.
Having said this, the answer will most likely lie somewhere between two to eight publications.
Many people argue that the quality of the venues is more important than the
This is 100% True. It’s much better to have one publication at a top-tier conference than to have dozens of publications on small workshops nobody is reading.
However, you need to have both quantity and quality to have a shot for those top-tier conferences with acceptance rates close to 20%. It’s not an either-or decision. You need to have it all.
Think about it: to get accepted on two top conferences (assuming a
Therefore, if you submit a research paper every two to three months, you are on the safe side and you will get accepted to two to four conferences during your research time. This will be enough (content-wise) to set up your first draft of the doctoral thesis. A few rounds with your supervisor and you are good to go! 😉