It looks like you're new here. If you want to get involved, click one of these buttons!
this is in chat because it's "half" related to Azimuth. Some of you may know the newly developing language Julia for scientific computing. One of the things in the "theory of technology" that seems well founded to me is the observation that a technology won't take off because it's just "a bit better" than what already exists, because any conversion has immediate costs that make it unwise to migrate just for a small benefit. A technology has to be "much better" (it's commony stated as "twice as good") in order to make converting existing stuff worthwhile. Julia is the first thing that I've seen in a while where I've felt it might be able to replace signficant amounts of Matlab/fortran/C++ (technical reason: it's actually being designed with JIT compilation in mind, whereas retrofitting a JIT to those languages runs into issues with their existing semantics.) It's also young enough that signficant changes in the language are still possible.
As such, I've been thinking about how to extent Julia (or, in the event the Julia guys don't want it, some other language) with array programming syntax. I've had to do this non-openly until I received written confirmation from my employer that this after-hours activity would be considered completely separate from anything I produce for them. That's finally come through, and so I'm going to start developing in the open. The repository is over at github. Note that it's not even functioning yet: after seeking some advice from Graham Jones about trees I think I've finally figured out a useful representation for loop nestings, but I still have to actually program that in. So it' s not remotely ready for actual use yet, but I'm putting it out there as part of doing open development (now that I'm finally able to do so).
(BTW: haven't mentioned the github repo being up on the Julia development mailinglist yet because I'd like to at least get the loop-nest tree search implemented before I get people looking at it, as they're more likely to persevere with it from that point.)