Hacker News Reader

Comment by hugofirth | Hacker News Reader

One reason I would love to see this succeed which has almost nothing to do with the particulars of the Frege project itself is the effect it would have on the Scala community.

There is a significant portion of the Scala community which desperately wants Scala to be the Haskell they can get their bosses buy in for. IMO the result is unsatisfactory, both for those people and the Scala community as a whole.

If a straight Haskell variant in the JVM ecosystem were to "take off" then Scala could be its own thing.

View on HN

So thats the reason most Scala code reads like badly written Haskell fanfic.

I think most scala code reads like badly written Java code, mostly because people use at as a "better Java". FP fanatics try to use it as Haskell for JVM, but they are definitely not the majority.

Most scala users have no f-ing clue what a monad is :)

To be fair, most programmers have no clue what a monad is. To understand what a monad is you first have to understand what a monad is. (No, that wasn't a typo.)

Ha true :)

I was making a jest to say that most scala programmers don't know anything about FP and just use it for the type inference and stick to OO practices, vars, etc.

No I think almost all Scala programmers understand the basics of FP.

They just don't understand / have much interest in the intermediate/advanced parts. I've seen lots of beginners take to for comprehensions, maps, filters etc pretty quickly.

The problem with Scala is that at an advanced level the code can become pretty unreadable.

To be fair, I'd hardly say that not knowing what a monad is = not knowing anything about FP.

I'm honestly curious why Kotlin hasn't gathered more steam amongst this (gigantic) group of people. It seems like there's definitely room for a "Java, but a bit nicer", but that's not what Scala is intended to be. Did Scala just start winning before Kotlin was around?

Did Scala just start winning before Kotlin was around?

I think so. By the time I heard of Kotlin, I mentally lumped it in the basket of "just another JVM language". I mean, there are so many now, and you already had Groovy, Clojure, Scala, JRuby, and Jython pretty well established, along with a couple of dozen other niche players. Then Kotlin comes along... I don't know about most people, but I never heard anything about Kotlin that was so compelling that I felt the need to go mess with it. I mean, why what instead of Nice, or Fantom, or Gosu, or Beanshell, or Mira, etc., etc...

Same thing for Ceylon. It looks like just another "also ran" to me. If the people behind these languages want people to use them, they are going to have to work hard to get the word out about whatever advantages (purported or real) they have.

I'll tell you why. Because right now Kotlin is the JVM language made by the largest company outside Java and Oracle. Well, technically, Ceylon is a Red Hat project, but Red Hat is a company that's basically built on lots of very loosely-related projects (Red Hat also fund JRuby, I believe), and one cannot say "Red Hat has Ceylon's back", while Kotlin certainly has JetBrains' back.

As a result, it is also the non-Java JVM language with the best IDE support, best Android support and best build-tool support.

Because right now Kotlin is the JVM language made by the largest company outside Java and Oracle

Truth be told, I don't really find that very compelling. I mean, do programming languages always need "big company" support to be successful? I don't know. So far Scala, Clojure and Groovy have done pretty well, with varying degrees of commercial support.

As a result, it is also the non-Java JVM language with the best IDE support, best Android support and best build-tool support.

I guess, but here's the thing... the languages I use now (mainly Groovy and Java) have at least good enough IDE support, Android support (ok, I don't really care about that) and build-tool support. So even if Kotlin is incrementally better, or hell, even dramatically better, that stuff doesn't do a lot to sway me to Kotlin.

That said, like most geeks, I like dabbling with new languages, and I am sure I'll try out Kotlin (and Ceylon) at some point. And maybe then one or the other will "wow" me. But for now, it isn't a high priority. And I expect a lot of other people are approaching it with a similar mindset.

> do programming languages always need "big company" support to be successful?

History shows that the answer to that is a resounding yes. Aside from a couple of scripting languages (BASIC and Python), there have been exactly zero successful mainstream languages without a large company behind them. Scala, Clojure and Groovy are doing fine, but their adoption is one or two orders of magnitude less than the mainstream languages (Java, C, C++, C#).

I think you're basically right, but the waters are murkier than you give them credit for. I don't really buy any definition of "successful mainstream languages" that doesn't include Ruby and PHP, and neither of those started with any sort of corporate backing, and still aren't tightly associated with any single "big company".

I think Ruby is too small to be considered an industry-wide success. PHP is indeed extremely successful, and should be added to my small list of scripting languages (in fact, it has been far more successful than Python): BASIC, PHP, Python. However, those exceptions are still only for scripting languages (even if you count Ruby), and none of the languages mentioned by the comment I responded to -- except Groovy -- falls in that category.

I think Ruby is too small to be considered an industry-wide success.

I'm not necessarily the biggest fan of the TIOBE index, but in this case I'll cite it, as it is "close enough" I think. Ruby is #13 right now, which isn't bad. And while I think TIOBE has some warts, I think it's safe to say that anything in the top 20 is fairly successful, and anything in the top 50 is "successful" to a certain degree.

http://www.tiobe.com/index.php/content/paperinfo/tpci/index....

This thread is a bit old for me to expect a response, but I'm always curious when I see people using this terminology: what purpose does the term "scripting language" serve? It doesn't seem to tell me anything about what a language is useful for, its runtime characteristics, or the style or philosophy it encourages.

> a large company behind them. Scala, Clojure and Groovy are doing fine

You do know VMware/Pivotal pulled their support for Groovy in March this year and no-one else stepped in, don't you?

Yes. None of them has a large company behind it, and none of them is within two orders-of-magnitude from the leading batch of languages (the leading pack has 5-10M developers each). Maybe Scala is, but even Scala certainly isn't within one order-of-magnitude.

None of them has a large company behind it, and none of them is within two orders-of-magnitude from the leading batch of languages (the leading pack has 5-10M developers each).

So? I'm not making a claim that a language needs adoption on that scale to be considered "successful". To me, Ruby, PHP, Python, Groovy, Scala, Clojure, etc. are all very much "successful". The languages I think of as being less so, would be things like Nice, Fantom, Frege, Forth, Modula-3, Rebol, etc.

You do know VMware/Pivotal pulled their support for Groovy in March this year and no-one else stepped in, don't you?

I do, but I also know that Groovy is becoming an ASF project[1]. So we'll see how it goes with volunteer support, and perhaps a few paid people here and there from companies that use Groovy.

[1]: http://incubator.apache.org/projects/groovy.html

> Groovy is becoming an ASF project

Groovy has lingering problems in migrating to builds.apache.org, see [1]. Some of the Groovy despots from Codehaus times (I wouldn't know which ones specifically) are keeping some computing machinery physically separated from the Apache infrastructure to run their own build processes, going against Apache guidelines. Some of them also have control over the groovy-lang.org domain (again, I don't know who). I suspect Groovy won't ever become an ASF project, but instead just sit it out in the incubation system until the former Codehaus Groovy despots get a better deal where they don't have to share their control democratically with the ASF. They managed to keep Groovy in the Java Community Process for 9 years before being booted out, all the while using the JSR-241 to promote themselves, so they'd think nothing of leeching on Apache's incubator for just as long before stirring up conflict later on to get booted out when it suits them.

BTW, that link [1] to Nabble's Groovy mailing list archive was recently redirected to an embedded view within the groovy-lang.org website where they can collect IP addresses, and all that implies -- another indictment of the Groovy despots longstanding intention to control and instead of sharing.

[1] http://groovy.329449.n5.nabble.com/Jenkins-Groovy-Apache-etc...

I don't see Kotlin as a "wow" language, I see it as a "this is sort of boring but pretty nice" language like Java and Go. Maybe Ceylon is similar, I dunno. It just strikes me that Go could use a stronger competitor than Java that runs on the JVM, but doesn't seem overly fancy (Scala, Clojure) or too dynamic (Groovy, JRuby, Clojure). Kotlin is the closest I've seen to that.

I completely agree. Kotlin is designed to be a modern Java, i.e. a blue-collar language that only adopts tried-and-true ideas.

> Did Scala just start winning before Kotlin was around?

Scala has about 2% mind share on the JVM after ten years in existence, I would say that not only is it not winning but its time has passed.

Ceylon is one year old, Kotlin is not even out yet, there is plenty of time for either of these two to gain some solid mind share.

Or maybe not. Maybe Java is going to reign supreme for quite a while. Whatever the replacement of Java will eventually be, I'm pretty sure it won't be Scala.

Kotlin is still in development, and the documentation is patchy. I couldn't recommend it for production usage just yet. However, having used a bit of it in an Android app, I think in a couple years it has a chance to be the canonical "java but better". At least I hope so.

The thing is that while Scala's main focus is not on being "a slightly nicer Java", it does that job fairly well.

There just doesn't seem to be a good point in using Kotlin, as you can tell from how fast the Kotlin proponent in this thread is changing the frame of reference.

- Higher-kinded types? Better compare with Haskell!

- Compilation speed? Let's pick Java, the language with the least useful typesystem!

- Popularity? Let's compare Scala with Java, but compare Kotlin to Scala!

- Commercial backing? Let's conveniently ignore that the language lives by the cross-subsidization of JetBrain's IDE business, instead of relying on the success of their language.

Point is, there is nothing that Kotlin does substantially better that would give it a niche between Java and Scala.

- Non-incremental compilation speed is not that much faster compared to Scala, and substantially slower than Java.

- Most of the things regularly used in Scala cannot be expressed in Kotlin, it only makes Java idioms slightly nicer to write down.

- Like in Scala, IDE support is ongoing work.

- Compared to both Java and Scala, the ecosystem isn't there.

Compared to Scala, Kotlin makes you pay 80% of the cost for 20% of the benefits.

I predict that future Java releases will keep cannibalizing Kotlin from the bottom, while Kotlin will fail to attract developers from more expressive languages.

>don't know anything about FP

Depends on how we define FP. Until 2010, when Haskell started to emerge as a HN fad, people were OK to define FP as what LISP programmers do, and didn't demand FP programmers know what a modad is, type theory or even use "immutability" everywhere...

If you read discussions from 2000-2005 for example, very few people define FP as "what Haskell does".

Has Haskell really been a fad here for five years? It feels like just a year or two ...

I think a lot of the problems with monads is that they're way too low level to really get a grip on what they're for --- it's like trying to explain algebra by talking about manipulating pixels on a screen.

Once I finally saw some of the things you can do with a monad rather than those useless examples based around Maybe, I had an 'aha!' moment as everything went click.

No sodding idea what a monoid is, though.

Ooo! Something I think I can answer! A monoid is just a thing that knows how to add to itself!

My understanding of a monad is that it isn't currying or recursion though.

I shouldn't even call it an understanding of monads. I've read at least 15+ explanations of what a monad is and still don't grok it.

Apologies in advance for the unsolicited monad tutorial (it's my turn to be that asshole)...

Firstly, saying "`X` is a monad" is the same as saying "class `X` implements the monad interface".

The monad interface is usually defined with `return` and `bind`, but I think it's more instructive to borrow the terminology from JavaScript's `Promise`...

* `Promise#resolve(value)` is the same as `return`, also known as `pure`. It just wraps the given value in a promise.

* `Promise#then(function)` combines both `bind` (when a new Promise is constructed and returned) and the functor method `map` (when a plain value is returned).

To provide a unified monad/functor interface for both `Promise` and `Array` (using `wrap` instead of `resolve):

    // `then` provides both `bind` and `map` depending on the return type of the given function:
    Promise.prototype.map = Promise.prototype.then 

    Promise.wrap = Promise.resolve

    // f maps elements to arrays of elements:
    Array.prototype.then = function (f) {
        return [].concat.apply ([], this.map(f))
    }

    Array.wrap = function (x) {
        return [x]
    }

There's no intrinsic value beyond that shared interface, it just allows you to write abstractions without knowing specifically what kind of monad you're dealing with (like the "do" syntax).

EDIT: BTW, I've purposely skipped a few things in this description, it's meant to be illustrative, not definitive...

Not quite ;) Semigroups are defined by their associative, binary operator of type a -> a -> a. Monoids are that plus an identity value of a.

I think using Haskell's type syntax is misleading in this context (and doesn't really imply the required associativity (much less commutativity if you wanted to consider Abelian monoids)).

The operation of a monoid maps from pairs of things to things. So in terms of types:

    <a,a> -> a

Or for some `twin` type constructor:

    twin a -> a

This is a bit more suggestive also in terms of F-algebras, where the operation has the following signature (f is a functor, or mappable container):

    f a -> a

There Haskell people go again =) You try to define something in simple normal terms, and off they go adding highly specific and technically correct words to what was supposed to be a simple explanation.

(I'm not a Haskell expert, I dabble, done CIS194, half of RWH, and I have no idea what you said -- which is a common problem I run into in the Haskell world, lots of super helpful people that have forgotten what its like to not speak their language)

A monoid is a type where we know how to combine two elements together. There are two parts: a function that combines and an identity for it. (The function is usually written as an infix operator to make it easier to read.)

For example, we can combine two integers by adding (+) with 0 as an identity. Or we could combine them by (*) with 1 as an identity.

Lists are a monoid because we can append two lists and the empty list is the identity.

The combining operator has to be associative and the identity has to be an identity for it. (Combining something with the identity gives you that original thing back.)

And that's all. There's just not much structure to them. In fact, there's so little structure that mathematicians don't care much about them. It's an exotic-sounding name for a fairly pedestrian concept.

But they're useful in programming. The main reason is that they're so ubiquitous: so many different things form monoids, often in different ways. This means we have a single interface applicable to almost every domain you can think of which is useful even if the interface doesn't tell us much.

They're also a good fit for parallelism. Because the combining function is associative, we can a bunch of combinations in any order we like, making it easy to spread them out over multiple threads. It very naturally captures the reduce in map reduce.

But mostly it's a convenient abstraction that's minimal and simple enough to pop up everywhere while still being useful.

Monoids are much easier. You just need a binary associative operator |+| (+,*, min, max all work for ints, ++ for lists/vectors, for instance) and some notion of an identity for the type s.t. forall a: A, a |+| identity === identity |+| a === a

They are HUGELY useful. One of the most useful is a "union" monoid definition for Maps. In Scala, it's written like this:

implicit def mapUnionMonoid[K, V: Semigroup]: Monoid[Map[K,V]]

So any map with Semigroup values is a monoid (Semigroups are monoids without the zero value). |+| under this definition will combine the Maps, but in hte case of collisions, |+| the colliding values together. This Monoid is the ABSOLUTE KING of aggregation. You can foldMap over lists and produce singleton Maps of the shape you want, and let the monoid instance aggregate for you.

Curious what you think of my monad tutorial! http://blog.reverberate.org/2015/08/monads-demystified.html

I'd sort of like to see 'return' explained earlier, rather than used without explanation in the IO monad example. Also it's weird to hear that a monad is a design pattern, and then hear about the monad deciding things; that implies something more concrete than just a design pattern. It's only vaguely clear that "the monad" in that sense is the specific choice of bind/return functions (if that's even correct!). But while it's a bit rough around the edges, I can see this becoming a really good explanation. It explains how the goals of monads relate the implementation, which seems to be the tricky part, in a way that makes sense to me.

   monads is that they're way too low level t

Monads are neither low- nor high-level. Monads are orthogonal to this. A monad (in the context of programming languages) is a generalised from of function composition. Instead of composing f : A --> B with g : B --> C yielding a function g;f : A --> C, monads compose functions f : A --> FB with g : B --> FC, yielding g;;f : A --> FC. Here F is come transformation on types. Monads connect the types FB and B in a canonical way. That's all.

I don't see why that should be the case though (apart from the prolification of bad, overwrought, explanations).

Monads are not that exotic. People present it as "deep math" but for a mathematician for example it's kids stuff at the level you need to understand them for Haskell et al.

It's like saying differential equations are some "deep voodoo math" (only monads are even simpler).

I think the bad, overwrought explanations are the whole reason.

Yes, they're much simpler than differential equations. But the difference is, differential equations are the simplest way to solve some difficult problems -- monads are often the gratuitously complex way to solve simple problems.

Monads are simple ways to solve simple problems.

Consider the Maybe monad - it encodes optional values in a simple way. The approach before this is to have something called "null" that would wreck your programs. Maybe monad is one of the best improvements to day-to-day programming tasks that I've personally experienced in my lifetime.

I agree, the concept of monads is very simple, actually trivial. It's just a generalised form of function composition.

But thinking monadically and abstracting monadically is extremely different from what programmers normally learn, for a start because important monads like state and exceptions are built-in features of most programming languages. Seeing that these things have a common pattern, and seeing that it may be worthwhile to abstract this common pattern takes a lot of time.

The mismatch between the utter simplicity of the concept of monads, and the complicated explanations one comes across doesn't help.

To be fair, (I think) you don't need to know category theory to successfully apply FP principles and/or use any functional language. I had a great (GREAT) success at using erlang not only not knowing what monad is, but even not knowing about existence of category theory at all, leave aside monads, monoids, functors, etc. You also don't need to know category theory to understand what monad is [0] This video is all you, as a practicing programmer, need to know about monads and it's just 1 hour long.

[0] https://www.youtube.com/watch?v=ZhuHCtR3xq8

Monads, Monoids, Applicatives, all very simple things. Most people make it seem difficult just to show off and feel as part of an elite. Which is a shame as it prevents wider usage.

James is one of the few people who don't make Monads into a mythical being.

http://james-iry.blogspot.de/2007/09/monads-are-elephants-pa...

There is nothing wrong with using Scala as a better Java. Indeed that's what I recommend to Scala beginners with a background in OO. That's the beauty of multi-paradigm languages: pick and choose the language subset that works for you.

Indeed I often program in a way that could be called "locally stateful, globally pure-functional". It's a good approach!

"locally stateful, globally pure-functional"

Can you expand on that, please? Because it seems to me that state fulness propagates up through the system. I can't imagine a system that was pure in the large but side-effecting in the small. I'd be interested to hear how that works...

One way is to use local mutable state within functions that are implemented so as to be not rely on external state, which are then indistinguishable to calling code from pure functions.

Yes, that's what I had in mind.

If only there was an effect system that could guarantee purity and at the same time not be in the way (i.e. allow full or Scala-like type inference). Then purity would be guaranteed by types, like it is in Haskell. (N.B. I'm not asking for Haskell's purity by default. I advocate impure as default, with a type guarantting purity.)

Out of interest - are you talking about the use of symbolic operators? Or something else (like the prevalence of the Typeclass pattern)?

Probably that Scala is extremely verbose and boiler platey compared to Haskell as soon as you start playing with higher kinds.

And when you use just regular Scala syntax, it's just pretty ugly code overall IMO (and one of the slowest compilers of the 21st century too).

The Haskelly parts of Scala are some of its best parts. It's always been a hybrid language, and it's always had a compromised design as part of that. But the result is wonderful, and either half alone would be diminished. And frankly any "boss-friendly Haskell" would have to have Scala-level Java interop, i.e. full support for inheritance, existentials and so on, at which point you end up going down the same path as Scala.

If you want Java you know where to find it. Likewise with Haskell. Scala draws from both and that's its greatest strength.

Hopefully any future languages that prioritize java interop will do things like implement java.util.List rather than insist on renaming "add" to "+=" along with ":+", "+:", "+", "++", "++:" so that I have to constantly call .asScala, .asJava on collections while passing them around.

Why not just import the implicit Java-Scala list conversions?

This obviously comes down to personal preference but I personally prefer the explicit conversions for the added clarity.

Disagree. The collections are different and behave differently (e.g. the Scala ones are immutable), they should look different. Explicit .asScala and .asJava make the intent clear. They should only be necessary at the boundary.

This is true - I love Scala for its hybrid nature. I'm just saying that I wish prominent members of its community would too. For example if you go on to the #scala IRC channel half the regulars on their will be complaining about how terrible Scala is ... For a variety of reasons that can be summed up as: its not Haskell.

If you spend a bit of time there it fades to background noise and #scala is actually a really friendly and helpful place - but "x is hard to do in Scala because Scala [sucks]" is an odd impression to give a newcomer.

There's an excellent implementation of OCaml for the JVM, which is, sadly, not well advertised (its author is a bit on the shy side): http://www.ocamljava.org/

One particular "annoying" feature of Scala is the Java interoperability. It's obviously great because you have a large quantity of tools usable from the Java world, but it leads to large quantities of time spent wrapping them into clean Scala interfaces (or dealing with unmelodious APIs, pick your curse).

I would love if I could avoid this with Fredge. But, in the end, I would probably be better off using Haskell directly.

I agree that when, in the process of learning Scala, you get beyond the "better Java" stage the interop can feel frustrating. On the other hand Scala gains so much from it. Its a huge net-positive IMHO.

As long as you are only bringing things in from Java-land, rather than going the other way though, the wrapping can be fairly minimal.

> I would probably be better off using Haskell directly.

This is certainly true. You can see Frege as a subset of Haskell 2010 plus some GHC extensions plus the native interface (i.e. the Java FFI).

Unless you really need the JVM, Haskell (i.e. GHC) is the probably the better choice.

Frege is just an offer for the minority that wants pure FP specifically on the JVM. Those people had no real choice until recently, given that CAL is dead, and E. Kmett's "ermine" not yet there.

This is about 50% of the pain of Scala for me. I want to like scala, and some of the concepts behind it are neat, but in reality the act of writing it with the java interop, opaque unrepeatable errors, and compiler slowness makes it a fantastically unpleasant experience.

I'm sharing the same thoughts. I'm relieved I'm not alone in this world :)

Reply on news.ycombinator.com