Why all the fuss with DSLs? Isn't it just "metalinguistic abstraction" as described in the SICP?

http://mitpress.mit.edu/sicp/full-text/book/book-Z-H-25.html#%_chap_4

by michaelneale • 19 years ago

4 6 19 years ago

mitpress.mit.edu

The current "fashions" around DSLs I find a little annoying, it seems like the idea has been around and in use for some time, as this shows. Is there something new and profound? It seems that DSLs taken out of the context of lisp are much much less powerful.

Its a good technique, "language oriented programming" it may be marketed as in a parallel universe.

mechanical_fish• 19 years ago

"Classic: a book which people praise and don't read."

http://www.twainquotes.com/Classic.html

In other words: Yes, in a world where every coder had an MIT degree, nobody would be especially impressed by DSLs.

Or maybe that wouldn't be enough. I've met several people who had SICP for their first-year programming class and did not like it at all.

eru2• 19 years ago

Real man don't eat quiche.

lisporama• 19 years ago

Aren't embedded languages (languages built "on lisp") superior to DSL's? Embedded languages can integrate with each other seamlessly if need be... and they can be manipulated in terms of the parent language.

eru2• 19 years ago

I agree. It should be much easier to build a decent embedded lingo.

AllMighty• 19 years ago

Having access to the full power of the parent language can be a problem. This point is well articulated here: http://www.w3.org/2001/tag/doc/leastPower.html

richcollins• 19 years ago

Why does the fact it appears in SICP make it any less interesting? Also, how often do LISPers use Macros for actual metalinguistic abstraction, rather than as a tool for code generation?

Why all the fuss with DSLs? Isn't it just "metalinguistic abstraction" as described in the SICP?

Related Stories