Comments on: Static Verification: An External DSL Advantage

By: Adia: A Week With Clojure And MongoDB « I am Zef

Adia: A Week With Clojure And MongoDB « I am Zef — Wed, 30 Dec 2009 16:43:13 +0000

[...] Clojure DSLs as flexible as external DSLs? Not really. First off, you’re locked into the Lisp syntax. Second, checking is rather [...]

By: Carson Gross

Carson Gross — Wed, 18 Nov 2009 02:59:47 +0000

There was once a Tower, built by man high into the sky…

By: Carson Gross

Carson Gross — Wed, 18 Nov 2009 01:59:47 +0000

There was once a Tower, built by man high into the sky…

By: Twitted by webstartupgroup

Twitted by webstartupgroup — Tue, 17 Nov 2009 23:34:43 +0000

[...] This post was Twitted by webstartupgroup [...]

By: How About A Nice Cup Of… « I am Zef

How About A Nice Cup Of… « I am Zef — Tue, 17 Nov 2009 20:07:42 +0000

[...] I thought, maybe it’s Rails people, that must be it. But, no, I don’t think so. Because after yesterday’s post, which is not really attacking Rails in particular, this [...]

By: Zef Hemel

Zef Hemel — Tue, 17 Nov 2009 13:29:20 +0000

In this post: http://zef.me/2371/when-scala-dsls-fail I show how Scala DSLs do this, but also where their limitations are. And indeed, the error messages are not good. This would also apply to Haskell.

And of course, it's a trade off.

By: donstewart

donstewart — Tue, 17 Nov 2009 13:25:33 +0000

> An external DSL compiler can detect mistakes early

A case you miss though: this propperty is also true of EDSLs in strongly statically typed languages (like Haskell) too. And it maybe be easier to embed the DSL's type system in the host language, than to design a coherent type system from scratch.

However, error messages are definitely far easier with DSLs.