Is it better to write a technical book instead of several research papers?

[A2A]

Answering as a computer science professor: Not if you want tenure!

The standard expectations for hiring and promoting faculty in most (if not all) fields of science and engineering include publishing research papers in peer-reviewed venues. In most fields that means refereed journals; in most subfields of computer science, that means competitive peer-reviewed conference proceedings. These venues are where the research community expect to see research results; publishing in these values is the easiest way to build both a public record of your research accomplishments and your reputation as a researcher.

(Expectations in other disciplines are different, especially in the humanities.)

It’s the easiest way not because it’s easy, but because that’s how everybody expects it to be done, so that’s where everyone’s attention already is. Hiring committees are going to look first at your publication record. Promotion committees are going to look first at your publication record. The people writing evaluation letters for your promotion case are going to look first at your publication record.

Rule Number One in dealing with committees that might evaluate you for anything: Thou Shalt Not Give the Committee an Excuse to Think. Publishing a book instead of several papers is non-standard,.and doing anything nonstandard gives the hiring/promotions committee an Excuse to Think. You will be held to a more stringent standard, because people are generally skeptical about unfamiliar things, and [whining] thinking is haaaaaaard. Some committee members will just want to reflexively reject anything “weird”, so other committee members will have argue more strongly to convince them.

Now, I’m not saying that it’s impossible to establish a strong enough reputation by writing a book to be hired and tenured. I’ve seen it work, but it’s rare. I don’t recommend the risk.