Cynbe's SML/NJ Page

I got interested in SML/NJ summer of 2004, which I spent surveying various functional and mostly-functional programming languages. I finally settled on SML/NJ as my favorite because:

• It is the de facto standard implementation in the field, supported by half a dozen leading figures in the field, so it is quite unlikely be orphaned and leave me high and dry.

• The maintainers are top people -- one can do worse than pick a platform based on who you get to work with!

• SML/NJ is the de facto standard guinea pig in the field, so there are lots of PhD thesis project type extensions to it lying around free for the taking. In particular, the CMU Venari Project implemented persistence, multiprocessing, concurrency, and transactions on top of SML/NJ -- which are all things I consider critically important.

• SML/NJ is based on an incremental compile-into-the-heap model much like that of Smalltalk, MUDs, X/emacs and such. Since I'd like to implement things like that, the architecture is a win from my point of view.

• SML/NJ has a liberal Berkeley-style license, so I'm free to hack up a storm without fear of lawyers and lawsuits. (I'd also be happy with GPL, but Ocaml's QPL is an albatross I'd rather not have hanging around my neck.)

To date, I've been busy learning the SML/NJ internals -- I've put together a page for anyone else crazy enough to diving into the internal plumbing.

Onward to Cynbe's SML/NJ Internals Page.
Back to Cynbe's ML Page.