Re: [Haskell-cafe] MultiCase alternative

On 16/06/2017, at 10:15 PM, Brandon Allbery wrote:

On Fri, Jun 16, 2017 at 6:12 AM, Baa wrote: Sure, but there is another point of view here. Alternative is to have flexible but elegant syntax embedded in language, instead of many-many extensions for any possibility - they "fragment" language. But of course this is controversial and ambiguous :-)

The active patterns of F# that you mentioned can be seen as a counterargument

There's an interesting irony here. I'm reading the paper "Extensible Pattern Matching via a Lightweight Language Extension" by Syme, Neverov, and Margetson. The authors propose active patterns as a way of allowing functional code to work with objects using pattern matching. The irony is that doing this clearly violates the Object-Oriented Programming principle "Tell, Don't Ask". That is, according to the OOP gurus, if you have some sort of if or case that classifies an object in some way, you are doing it wrong. That variant- dependent behaviour belongs in the object. I'm agnostic about most OO dogma, so don't take this as my claim.