Could APL be implemented in Scala as a DSL? - scala

There is a old computer language called APL. Could this be implemented in Scala as a DSL?
http://en.wikipedia.org/wiki/APL_%28programming_language%29

Someone could probably give a better answer than this, but this is my initial thought:
A Scala DSL should in theory be able to implement any programming language because it could build up an arbitrary structure representing the syntax, and then evaluate that.
A Scala DSL could not exactly replicate APL syntax for many reasons, one of which is that
'single quotes'
can denote a string in APL, but not in Scala. Also (from the wikipedia page)
×/2 3 4
wouldn't be valid Scala.
I don't know how close you could get, though...

A Javascript implementation exists here: https://github.com/ngn/apl

Related

How to avoid writing confusing DSLs in Scala

I've read comments stating that Scala's flexibility makes it easy for developers to write DSLs that are difficult to understand and reason about.
DSLs are possible because
we can sometimes omit . and parentheses (e.g. List(1) map println)
we can sometimes interchange () and {}
we have implicit values, parameters, and classes (also conversions, which are now discouraged)
there is a relatively small number of reserved symbols in the language (e.g. I can define + for my class)
and possibly other language features.
How can I avoid writing confusing DSLs ... what are the common antipatterns? Where is a DSL not appropriate?
Whenever you create DSL of your own you're embedding new language into Scala, which is not standard, so it doesn't follow standard code guides, conventions, etc.
I would say it's nothing wrong with adding new DSL as long you add proper documentation, explain the purpose of creating it and add examples of usage. If you feel adding new DSL would increase readability of your code, just go for it, but remember that whenever anyone encounters your DSL and it won't be documented enough, they will be very confused.
A good example of well-documented and serving good purpose DSL would be matchers of scalatest or Scala duration.

Separation of concerns

I am trying to learn functional programming in scala with book FPiS in chapter 5 the author mentions:
A major theme in functional programming is separation of concerns and
seperating program description from evaluation.
What does it mean? Could someone give an example?
Here I provided an example of implementing a tail recursion manually. Tail recursion - Scala (any language else)
It is an example of separation of algorithm description and evaluation.
Recursive trait describes only one iteration of some recursive algorithm.
Method interpret knows nothing about algorithm's logic and just runs it until it is finished.
For example you can introduce a delay between iteration or limit the number of iteration without changing the algorithm described in Recursive.
The key to this is in the phrase
seperating program description from evaluation
An example is using a DSL represented by an ADT (that represents the grammar of your DSL) and an interpreter. Because other people are probably better than me in describing this in detail, I'll just link to an example here: http://typelevel.org/cats/datatypes/freemonad.html
That one uses free monads which are a somewhat hot topic currently but demonstrate very good what you're asking for in my opinion.

Is there an equivalent of Haskell's CHP for Scala?

Clojure has this amazing library implementing Tony Hoare's Communicating Sequential Processes called core.async.
Haskell appears to have an equivalent called chp. (Not sure if it compiles under GHC 7.8).
My question is Is there an equivalent of Haskell's CHP for Scala?
You should have a look at this:
https://groups.google.com/forum/#!msg/scala-user/NljrQ4Mc-aI/3ISm68sqLNAJ
It provides a really interesting list of alternatives for doing CSP in Scala and doesn't recommend JCSP since it's development apparently stopped in 2011.
It also talks about a really interesting paper written by Andrew Bate at Oxford that describes a DSL in Scala for CSP but that whose implementation hasn't been open sourced.
It finally describes Quasar that seems the best alternative. If you're interested in Quasar, this post gives a good description of how Quasar works.

What parts of the Java ecosystem and language should a developer learn to get the most out of Scala?

Many of the available resources for learning Scala assume some background in Java. This can prove challenging for someone who is trying to learn Scala with no Java background.
What are some Java-isms a new Scala developer should know about as they learn the language?
For example, it's useful to know what a CLASSPATH is, what the java command line options are, etc...
That's a really great question! I've never thought about people learning Java just so they have it easier to learn Scala...
Apart from all the basics like for loops and such, learning Java Generics can be really helpful. The Scala equivalent is much more potent (and much harder to understand) than Java Generics. You might want to try to figure out where the limits of Java Generics are, and then in which cases Scala's type constructors can be used to overcome those limitations. At the more basic level, it is important to know why Generics are necessary, and how Java is a strongly typed language.
Java allows you to have multiple constructors for one class. This knowledge will be of no use when you learn Scala, because Scala has another way that allows you to offer several methods to create instances of a class. So, you'd rather not have a deep look into this Java concept.
Here are some concepts that differ very strongly between Java and Scala. So, if you learn the Java concepts and then later on want to learn the equivalent in Scala, you should be aware that the Scala equivalent differs so greatly from the Java version that a typical Java developer will have some difficulty to adapt to the Scala way of thinking. Still, it usually helps to first get used to the Java way, because it is usually simpler and easier to learn. I personally prefer to think of Java as the introductory course, and Scala is the pro version.
Java mutable collection concept vs. Scala mutable/immutable differentiation
static methods (Java) vs. singleton objects (Scala)
for loops
Java return statement vs. Scala functional style ("every expression returns a value")
Java's use of null for "no value" vs. Scala's more explicit Option type
imports
Java's switch vs. Scala's match
And here is a list of stuff that you will probably use from the Java standard library, even if you develop in Scala:
IO
GUI (Scala has a wrapper for Swing, but hey)
URLs, URIs, files
date
timers
And finally, some of Scala's features that have no direct equivalent in Java or the Java standard library:
operator overloading
implicits and implicit conversions
multiple argument lists / currying
anonymous functions / functions as values
actors
streams
Scala pattern matching (which rocks)
traits
type inference
for comprehensions
awesome collection operations like fold or map
Of course, all the lists are incomplete. That's just my view on what is important. I hope it helps.
And, by the way: You should definitely know about the class path and other JVM basics.
The standard library, above all else, because that's what Scala has most in common with Java.
You should also get a basic idea of Java's syntax, because a lot of books end up comparing something in Scala to something in Java. But other than the platform and some of the library, they're totally distinctive languages.
There are a few trivial conventions passed from one to the other (like command line options), but as you read books and tutorials on Scala you should pick those up as you go regardless of previous Java experience.
The serie "Scala for Java Refugees" can gives some indications on typical Java topics you are supposed to know and how they translate into Scala.
For instance, the very basic main() Java function which translate into the Application trait, once considered harmful, and now improved (for Scala 2.9 anyway).

What are the key differences between Scala and Groovy? [closed]

Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 7 years ago.
Improve this question
On the surface Groovy and Scala look pretty similar, aside from Scala being statically typed, and Groovy dynamic.
What are the other key differences, and advantages each have over the other?
How similar are they really?
Is there competition between the two?
If so, who do you think will win in the long run?
They're both object oriented languages for the JVM that have lambdas and closures and interoperate with Java. Other than that, they're extremely different.
Groovy is a "dynamic" language in not only the sense that it is dynamically typed but that it supports dynamic meta-programming.
Scala is a "static" language in that it is statically typed and has virtually no dynamic meta-programming beyond the awkward stuff you can do in Java. Note, Scala's static type system is substantially more uniform and sophisticated than Java's.
Groovy is syntactically influenced by Java but semantically influenced more by languages like Ruby.
Scala is syntactically influenced by both Ruby and Java. It is semantically influenced more by Java, SML, Haskell, and a very obscure OO language called gBeta.
Groovy has "accidental" multiple dispatch due to the way it handles Java overloading.
Scala is single dispatch only, but has SML inspired pattern matching to deal with some of the same kinds of problems that multiple dispatch is meant to handle. However, where multiple dispatch can only dispatch on runtime type, Scala's pattern matching can dispatch on runtime types, values, or both. Pattern matching also includes syntactically pleasant variable binding. It's hard to overstress how pleasant this single feature alone makes programming in Scala.
Both Scala and Groovy support a form of multiple inheritance with mixins (though Scala calls them traits).
Scala supports both partial function application and currying at the language level, Groovy has an awkward "curry" method for doing partial function application.
Scala does direct tail recursion optimization. I don't believe Groovy does. That's important in functional programming but less important in imperative programming.
Both Scala and Groovy are eagerly evaluated by default. However, Scala supports call-by-name parameters. Groovy does not - call-by-name must be emulated with closures.
Scala has "for comprehensions", a generalization of list comprehensions found in other languages (technically they're monad comprehensions plus a bit - somewhere between Haskell's do and C#'s LINQ).
Scala has no concept of "static" fields, inner classes, methods, etc - it uses singleton objects instead. Groovy uses the static concept.
Scala does not have built in selection of arithmetic operators in quite the way that Groovy does. In Scala you can name methods very flexibly.
Groovy has the elvis operator for dealing with null. Scala programmers prefer to use Option types to using null, but it's easy to write an elvis operator in Scala if you want to.
Finally, there are lies, there are damn lies, and then there are benchmarks. The computer language benchmarks game ranks Scala as being between substantially faster than Groovy (ranging from twice to 93 times as fast) while retaining roughly the same source size. benchmarks.
I'm sure there are many, many differences that I haven't covered. But hopefully this gives you a gist.
Is there a competition between them? Yes, of course, but not as much as you might think. Groovy's real competition is JRuby and Jython.
Who's going to win? My crystal ball is as cracked as anybody else's.
scala is meant to be an oo/functional hybrid language and is very well planned and designed. groovy is more like a set of enhancements that many people would love to use in java.
i took a closer look at both, so i can tell :)
neither of them is better or worse than the other. groovy is very good at meta-programming, scala is very good at everything that does not need meta-programming, so...i tend to use both.
Scala has Actors, which make concurrency much easier to implement. And Traits which give true, typesafe multiple inheritance.
You've hit the nail on the head with the static and dynamic typing. Both are part of the new generation of dynamic languages, with closures, lambda expressions, and so on. There are a handful of syntactic differences between the two as well, but functionally, I don't see a huge difference between Groovy and Scala.
Scala implements Lists a bit differently; in Groovy, pretty much everything is an instance of java.util.List, whereas Scala uses both Lists and primitive arrays. Groovy has (I think) better string interpolation.
Scala is faster, it seems, but the Groovy folks are really pushing performance for the 2.0 release. 1.6 gave a huge leap in speed over the 1.5 series.
I don't think that either language will really 'win', as they target two different classes of problems. Scala is a high-performance language that is very Java-like without having quite the same level of boilerplate as Java. Groovy is for rapid prototyping and development, where speed is less important than the time it takes for programmers to implement the code.
Scala has a much steeper learning curve than Groovy. Scala has much more support for functional programming with its pattern matching and tail based recursion, meaning more tools for pure FP.
Scala also has dynamica compilation and I have done it using twitter eval lib (https://github.com/twitter/util ). I kept scala code in a flat file(without any extension) and using eval created scala class at run time.
I would say scala is meta programming and has feature of dynamic complication