I came to the field of programming language research as an outsider. Though we had an active researcher in our faculty (he has since retired), for various reasons he was never my mentor, so I never got personal introductions nor did I receive much oral wisdom from an elder in the field. Instead, I immersed myself in the literature. Eventually, I got good enough to write a reasonably good licentiate thesis, which in turn led me to spend three months visiting one of the external examiners, Stefan Hanenberg. From him, I got some of the inside story, and the world looked much different. Of course, he is a minority voice in the field, but every participant has a unique point of view anyway. The thought I am writing about here crystallised for me immediately: too much of the field’s development happens off the record!
On Wednesday, I presented my essay “Concept analysis in programming language research: Done well it is all right” [ACM DL] [Author’s PDF] [presentation slides (PDF)] at SPLASH Onward here in Vancouver. I told some of my story there; the session chair Robert Biddle expanded on it and made a forceful point, which I am repeating and expanding on here now.
The discussions that lead to significant developments in the field must happen on the record! It is fine to talk with friends and colleagues in pubs and at lunch (or wherever), but if the discussion leads to a concrete proposal that would affect the field either substantially (in terms of, for example, conceptual developments), the issue should be written up and published in a publication of record, and sufficient time should be allowed (if possible) for contrary and refinement views to be similarly published on the record.
The reason for this is, on the one hand, the empowerment of the community fringe, who does not have the opportunity to participate in off the record discussions, and, on the other hand, the creation of a full record for the future generations of researchers so that they can read up and learn about why things are the way they are.
Concept analysis, as I proposed it in my essay, is one way of proceeding with this on-the-record development of the field in terms of conceptual issues. Too often it appears to an outsider that things just appear out of thin air. Instead, any conceptual developments should be argued for in the literature!
I think the field would benefit enormously if we stopped thinking of research publication as the accumulation of facts (or the completion of a grand theory), and instead took a page from the humanities and the social scientists: for them, scholarly publications form a grand – multicentennial – discussion where individual researchers listen for a while, then start participating for a couple of decades, and then go away, while others take their place. This viewpoint has a side-effect of creating a fuller historical record, but it also places more responsibility on the reader: you have to listen for a while to catch the import of what you are reading, instead of grabbing a paper here or a paper there and taking them to be the gospel.
It would also allow putting more of the development of the field on the record.