On Mon, Mar 31, 2014 at 12:59 PM, Dan Bron <[email protected]> wrote:
> So the analogy in J to searching for types in statically-typed languages
> would be to search for sequences of words?

Hmm... there's quite a lot one could do with a J search engine.

The equivalent J search engine would probably be a dictionary search,
based on type (noun, verb, adverb, conjunction) and maybe result rank.
You might also provide example an example right argument (or left and
right arguments) to constrain the search to routines within that
domain.

> As in, just as you could search as Haskell codebase for functions of type
> "int" (primitive) or "int -> (int, [char])" (compound) or "data Suit =
> "Club | Diamond | Heart | Spade" (named) or whatever, you could search a J
> codebase for functions like "+" (primitive) or "+/@%" (compound) or "avg
> =: +/ % #" (named) or whatever?

It doesn't look to me as if the Haskell codebase search finds
application code, though I suppose it could.

Actually, an interesting thing for a J codebase search would be some
way of finding special code (perhaps based on primitives involved). Or
maybe that should somehow be incorporated into
http://www.jsoftware.com/jwiki/NuVoc? Perhaps "special code" links
from each primitive page where the primitive is involved in special
code?

> Is that what you mean by drawing an equivalence between "static type
> searching" and "syntactic searching"?

Roughly speaking, yes.

Thanks,

-- 
Raul
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to