This should be of great interest to anybody working with high volume sensor data in IOT field. I will be checking it out very soon (on a few deadlines now so cant afford the diversion). Anything that is fast and simple for massive volumes of small record types should always be of great interest to IOT/WOT folks.<p>I agree with the comments about KDB/Q - tried to look at it but could never understand it. Maybe cos im not fulltime dev, nor in that field, but KDB just always becomes too hard. AT least hobbes looks like i can work it into C++. Looks at first 'parsing' to be a case of right tool for the job of working on high volume discrete data which one would expect from an investment bank.<p>Kudos to the bank for releasing this part of their secret-sauce.
All embedable languages should come with their own header parser/code generator to save us from the hassle of generating a bunch of class wrappers and boilerplate registration code. I think this high up in the list of concerns in a professional setting.
This looks like a typed take on KDB/Q, something that is long overdue! The key to making this work is structural typing of rows (row polymorphism / extensible records). I'd be interested to see more details on how they tackled this.
I'd love to see more about this, seems very interesting language. Something to explain how/where this is/could be used, maybe few more complete examples to show the language etc. Also some notes about performance would be great, I guess it is reasonably fast by the looks of it, but that is pretty vague.
Is this a <i>"pure"</i> language like Haskell? Or it's more like OCaml?<p>And, even if it's not pure, is there any syntactic sugar for monads?<p>Anyway, this looks <i>awesome,</i> especially with easy C++ interfacing that seems to be there: a system where you can have machine learning (anything from bayesian to deep nns) code in C++ and business logic in something Haskell-like is the stuff of wet dreams...
Can someone explain variants vs sums vs tuples vs records? This sentence from the README confuses me: "We can combine types with variants or sums (the "nameless" form of variants, as tuples are to records)."
> <i>you will need LLVM 3.3 or later</i><p>Though it doesn't build with 4.0. (And is now in the AUR.)<p>Nice to see first-class parsing in a language.