Date: Fri, 7 Nov 1997 06:31:08 -0500 (EST) Message-Id: <199711071131.GAA18773@locke.ccil.org> Reply-To: David Barton Sender: Lojban list From: David Barton Subject: Re: 'your will' as sumti X-To: AshleyB@HALCYON.COM X-cc: LOJBAN%CUVMB.bitnet@UCCVM.NYU.EDU To: John Cowan In-Reply-To: <199711070942.EAA16014@severn.wash.inmet.com> (message from Ashley Yakeley on Fri, 7 Nov 1997 01:42:15 -0800) X-Mozilla-Status: 0011 Content-Length: 1483 X-From-Space-Date: Fri Nov 7 06:31:12 1997 X-From-Space-Address: LOJBAN@CUVMB.CC.COLUMBIA.EDU Ashley Yakeley writes: I think this called 'currying' in functional calculus... anyone care to comment? Hmmm. I don't think so. "Currying" in a functional language refers to the fact that, because functions are first class values (the same class of values as numbers and characters), a function can be formed from a multi-parameter function by providing just one of the arguments (or two, or any number less than the full number called for). Thus, if "add" takes two parameters, (add 2 2 being equal to four), then "add 2" is a function that takes any number and adds two to it. "add 2" is said to be a "curried" application of "add", and the general process is called "currying" (after Haskell Curry, who worked out the type system of the lambda calculus and who codified this concept very explicitly in his theory). I can then set "f = add 2", and later say "f 3" which produces the answer 5. It seems to me that the example to which you were referring was simply the ability to leave out the ending terminator, simply because no ambiguity results. This is a different concept. I see no ability in Lojban to set a pro-sumti to a partial sumti (ko'a equal to "mi klama", say) and then to "fill in" the left out positions later (ko'a le zarci"). This would be currying. Dave Barton <*> dlb@intermetrics.com )0( http://www.intermetrics.com/~dlb