Pacific Northwest Scala 2013 We're Doing It All Wrong by Paul Phillips

Correctness is the dog. Everything else is tail. Don't know who said that first, might have been this guy this video

>Edward Kmett is a Haskell evangelist Really? Could you point me to some of his evangelizing? Or did you mean "he uses haskell and is quite happy with it"? Because that isn't the same thing. >inconsistencies that the average programmer will never even realize exist. Do you honestly believe that? this video >Scala that only when you leave it's confines do you realize, shit, I miss But the very point Edward makes is that you don't give up those things when you leave scala. You only give up those things if you leave scala for something worse. You could leave for something better, keep all those things, and get tons of important things that you also won't realize you are missing until you get used to them and then try to live without them.

Very True. I'm pretty sure that quote wasn't meant to be a critique of the Groovy language, just an expression of interest in Scala. Yet its brought out every time somebody wants to dismiss Groovy. Compare that quote to this[1] presentation where Paul Phillips lays in to scala for 50 minutes. Even when somebody does criticize their own creation I doubt they mean "Hey everybody, that thing I spent a bunch of time on? Throw it all in the garbage, its pointless". [1] this video

Some context - the lead developer on the Scala Compiler, Paul Phillips, went slightly AWOL last year, and presented the talk "We're doing it all wrong" this video link He still seems slightly burnt by the whole thing: link Other critiques of Scala collections have been made with the Typelevel library link which aims to fix these issues. This post, 'Next Steps', seems to have come out of Martin Odersky's response to the situation. I commend him for taking practical and thought out action.

Paul, why do you say it's wrong that Float(Long.max - Int.max) is equal to Float(Long.max)? Float has less than 32 significant binary digits, and Long.max - Int.max has the same MSBs as Long.max. You would need to use Double to see the difference. I'm not sure what implicit conversions/promotions have to do with that - you want a warning or error when going from Long to Float? That would be reasonable. You can even lose significant digits between Int->Float and Long->Double even though they have the same total number of bits, so you could require those to be explicit. (from a slide in this video )

a) Martin Odersky has admitted as much[1]. b) The compiler internals are reportedly a mess[2]--likely due at least in part to the extreme number of different syntax variants. c) I find rule 4.1 in OP interesting ("SHOULD avoid concurrency like the plague it is"). According to Odersky's Coursera course[3], one of the main advantages of functional programming as enabled by Scala is easy, safer concurrency (reduced/absent mutable shared state). Yet here we see a presumably seasoned Scala veteran basically saying that this is a failure and we should avoid concurrency whenever possible because it results in too many bugs. 1. link 2. this video 3. link

Oh damn, I didn't know they were the same person. All I knew about him beforehand is from watching this video[1], but it's weird to read about a person and then find out that they also were the same person that you knew from somewhere else. [1] (How Scala [Typesafe] is doing "everything wrong") this video

Really good points. I expect you're right about the expressiveness, although I don't really know exactly what you mean by this. The "buffet of abstractions" available in Scala actually put me off it. For me, there are too many options available to expect to have readable, maintainable code at the end of it all, especially if you work in a team which regularly creates almost wilfully hard-to-comprehend java. This video is 2 years old now, but I think it marked the point when I stopped investigating scala. this video

Scala's new compiler (dotty) also tries to go that way (first of all using streaming AST-s for better cache locality). The problem is, that usually this "keep state and just percolate changes" is easier said than done. But we're getting there. See also: this video (scalac performance) this video (what do we really need from "computing science" to do programming) link

A bit tangential, but - an ex-Scala core contributor's lament about why compilers have to be incremental: this video - a Rust core team member talking about responsive compilers: link (basically the "low-level" implementation of the talk at the previous link)