Friday, September 25, 2009

Why design a new programming language?

So why would anyone want to design a new programming language? For some of us who have the bug, it is the ultimate design project. Imagine actually creating the language in which you can express yourself. But there is another reason. I have been in the software business for over 40 years, and despite everything that might have been said to the contrary, I still believe that a well-designed programming language can result in more productive programmers building higher quality software. In the particular area of high-integrity software, including both safety-critical software and high-security software, there is all the more reason to use the very best programming language you can, because the problems you are trying to solve and the level of quality required is at the very limits of what can be accomplished.

This new language is meant to address the goals of producing inherently safe and secure software, while taking advantage of the wider availability of true parallel processing in the form of multi-core chips. It is intended to promote a formal approach to software, where the program text includes pre- and postconditions, liberal use of assertions and invariants, etc., with tool-supported proof of correctness with respect to the formal annotations.

The language is tentatively named ParaSail, for Parallel Specification and Implementation Language. I would have spelled it "ParaSAIL" but for the danger of confusion with the original Stanford AI Language, "SAIL," and its more modern follow-on "MAINSAIL" (for Machine Independent SAIL). I don't mind making the connection with SAIL, as it was a very interesting language in its day, and MAINSAIL remains worth a look today. ParaSail is a completely new language, but it steals liberally from other programming languages, including the ML series (SML, CAML, OCAML, etc.), the Algol/Pascal family (Algol, Pascal, Ada, Modula, Eiffel, Oberon, etc.), the C family (C, C++, Java, C#), and the region-based languages (especially Cyclone). Perhaps one significant deviation from the excellent baseline established by SAIL, ML, Eiffel, Java, etc. is that ParaSail is intended to avoid "fine-granule" garbage collection in favor of stack and region-based storage management.

So why a blog? I guess I'll leave that to the next post.


  1. NOTE: If you want to post a comment using the Firefox browser, you will have to enable 3rd-party cookies. I agree that's a bit annoying!

  2. Have you looked at the language Agda?

  3. Thanks for the pointer. One issue with languages like Agda is whether they would be accessible to the typical developer of safety-critical systems. It is a balancing act to produce something that is innovative and pushes the state of the art while still being accessible to the relevant developer community.