The Pluvo Release

On the 1st July I released Pluvo, the scripting language that I've been hacking on since the middle of May. I'd effectively stopped working on it, but decided that rather than have it just lying around and not doing anything it'd be better if I released it. I didn't expect to get much feedback, but at least it would be out there to find for those that were looking, and the handful people that did actually know about it could get a proper look.

But of course I released it to the Swhack Mailing list and miscoranda, and it seems that it's from the latter that Danny Ayers picked it up and published a link to it on Lambda the Ultimate, which ensured a steady flow of links and a fair few peripheral bloggings and comments. One surprise is that the huge majority of the feedback has been positive, and much of it has been exceedingly so. Only once or twice has it been asked what the point and justification is, which is excellent since I don't think it really needs justifing for a start, and the point of it is fairly clearly evident from the structure. Most people seem to have picked that up, and for those that haven't I've now added an FAQ on it.

My earlier natural language design essay came hot on the heels of the initial Pluvo development, and the mix of the two have meant that language design on all fronts has been something that I've been very much focussed on at the moment. As for the future of Pluvo, I've had a lot of feedback, not least from Joe Geldart, and it's been almost too much to take on board. The biggest problem is mediating the features that ought to go into the language whilst placating the people supporting those that don't make it. When a language is in its very early stages of development, that becomes exceptionally tricky. I'm glad that I waited for it to be fairly mature in its structure before releasing it otherwise the problem would've been multifold confounded. As it is, Joe thinks it's a bit semantically ill defined, so he's planning on possibly releasing a Haskell version to shore that up a bit.

For Python, Perl, and Ruby the benevolent dictator approach seems to have worked marvellous grand. On the other hand, those same dictators were excellent programmers and could better see those languages through the early days. Now we have those languages to leverage off; and so in turn future languages will gain leverage off Perl6 and so on, though there it seems that everything is just expected to be a dialect of Parrot for some time. I'm not sure that the practicality of that will work out quite as planned, because the semantics of a language are actually quite tied to its syntax—see for example the problems that have been encountered just in trying to get Python hosted on Parrot.

Anyway, my aim for Pluvo still remains the same: to have a simple syntax on top of which a great deal is possible, to be as steady as bash but as flexible as Python. And to integrate a lot of other ideas, e.g. prototypical object orientation and hybrid typing, along the way. So the medium term aim at least is to add it to that list of currently two languages that I always turn to when I want to implement something at the moment, and unfortunately I'm wondering if I'll really turn to Pluvo until it's implemented in a more solid and widely available language. Much of what I do, for example, scriptwise is CGI, and if the Haskell implementation is to be the reference implementation then that's great but Haskell is hardly as easy to mess around with as Python.

So it would be good if I could drag more people into getting interested in it, but then the language is bound to swerve and run all over the place if I ask for that much commitment from people, and I'm not all that bothered anyway. It was very fun to spec out, and it's there as a framework to absorb future ideas that I have, and people are interested in it and I'll knock about ideas with them too as I have been doing already. Not so much of a snowball process, but at least it should gradually keep moving forward, and then who knows? It took Python and Perl many many years to get to any sort of maturity anyway, and I think that that situation is quite comparable to Pluvo at the moment, though I don't expect Pluvo to become anything like a major player in the realm of programming langauges. It would be nice if it because a beloved curiousity though, and again something that I can use when the occasion is fitting for it.