Lambda the Ultimate

I thought this might be interesting to discuss. You see, I work on the TUNES project as the lead research and development programmer (in the last year or so). Unfortunately, this project suffers from too broad a scope and some rather questionable views pervading the entire set of documentation. This project has been around since October1994, with a mailing list composed mostly of people without the time and understanding combined to give the project the progress it justly (in my opinion) deserves. I wrote an open letter concerning the state of TUNES the other day, which addresses several points. Now, two projects have been started, Slate and Arrow, which have terrible documentation I admit, but this is due to my current complete lack of time, and the lack of volunteers and vision in TUNES. Currently I am working on an alpha release of Arrow over Squeak and there are a few others who might join efforts, but this still doesn't amass the kind and concentration of effort we need (such as site organization). So my question is what do you suggest?

[later]...but of course, I don't. Meantime, you might need this (according to the link above, it's background info).

[later still] Using category theory would tidy up your discussion of arrows (sorry if I missed a mention of this). It's very difficult to tell whether this makes sense - it's worryingly close to something out of Sokal & Bricmont in places, but maybe that's just my restricted upbringing showing through.

I now see why the Arrow reference wasn't in the original post; it's not a language as much as an application (I think).

I wrote that paper while at sea for six months, and it shows. My state of formal knowledge on it is much greater now, including Category theory, Lattice theory, substructural logics, and hundreds of programming language concepts that apply in certain contexts. The paper needs a major overhaul, obviously, but I am concentrating on producing a piece of software that demonstrates what I intend and letting that speak for itself and essentially be a basis for the revised paper. However, the distinction between language and application doesn't quite apply to what I'm working on, although there are few better metaphors. I am definitely interested in using Arrow for a different kind of meta-programming than is usually done, but I'm interested in doing it without a programming language as such.

The software that is publically available titled Arrow is just a small part of what I am working on, and even the semantics are not really representative yet.