Syntax matters. People have been writing new programming languages to do this and that, but the one thing that the programming languages community has yet to address is syntax. Semantics are very important, but writing code in a language with a crappy syntax is its own special hell. This problem needs solving.
We can argue about Python vs Perl vs Ruby vs Lisp all day, but those arguments seem pretty trivial compared to the fact that people are still programming in FORTRAN, COBOL, BASIC, and IDL. Those languages have a syntax that, I contend, is actually verifiably crappy. They promote solution methods that break when touched funny, are hard to read, and usually contain subtle bugs. They make easy things look hard while still keeping hard things hard. Perhaps we should just say that if your language is a word in all caps, then it sucks. This would eliminate C, but not Objective-C. And it would eliminate C++, which pretty much does need eliminating.
Much like we currently have an active research community in semantics – type theories, advanced resolution methods, and the like – we need an active research community in syntax. Which means that we need to have some methods of testing which syntax is better. Because this is a problem with a large psych component, the data will necessarily be fuzzier and less conclusive. But currently nobody is doing anything about this. I know I don’t want to – HCI just isn’t that enjoyable at the end of the day (sorry Fiona, Isha, Tim, and many others). But perhaps it is something the HCI community should take up. They are in many resepcts ideally suited to begin this task, and it would let them publish their papers somewhere other than SIGCHI, which has too many submissions and is essentially a broken conference in which sexy demos drive out good science.
