Tony Morris

@takeoutweight http://hackage.haskell.org/package/mtl-1.1.0.2/docs/src/Control-Monad-State-Lazy.html#State … operates only on persistent data structures, though. so it can't be the defn you're looking for.

@mwotton I'm not happy w it but being charitable as it admits State but rejects ML ref cells. 'purity' is going on the buzzword list for now

@takeoutweight @mwotton That many people misuse a term, as you are doing, is not reasonable permission to reject its useful context.

@dibblego @mwotton What's the reference for the proper usage of the term? It seems to be used mostly as branding like "OO"

@takeoutweight @mwotton An expression is pure if equational reasoning is preserved. Now show me where the State monad fails this.

@dibblego @mwotton The point is that you can give monadic semantics to a language with mutable ref cells, so the word pure is meaningless

@takeoutweight @mwotton You are now trying to describe "languages" as pure? Yes, another futile exercise in unicornery.

@dibblego @mwotton Osaki uses the term pure to describe languages. Anyway I'm looking for a peer reviewed ref for proper usage if one exists