home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.lang.apl
- Path: sparky!uunet!cs.utexas.edu!qt.cs.utexas.edu!yale.edu!yale!gumby!ursa!ursa!abalje47
- From: abalje47@ursa.calvin.edu (Alan Baljeu)
- Subject: Re: Surrogate arguments in J 3.0
- Message-ID: <abalje47.727733817@ursa>
- Sender: news@calvin.edu
- Organization: Calvin College
- References: <qmdbms.727703760@gsusgi1.gsu.edu>
- Date: Fri, 22 Jan 1993 20:16:57 GMT
- Lines: 15
-
- Brian Schott writes:
-
- >I have been reading Alexei Skurihin's article "Neural networks in J" from
- >the Russian APL92 Conference. His code does NOT contain the surrogate
- >arguments x. and y. presumably because early versions of J accomplished
- >this differently: How?
-
- I believe this hasn't changed. However, it is possible in J to take a function
- using x. and y. and transform it so that it no longer contains any x.'s or y.'s
- This produces what is known as a pure, or tacit function. See APL91 for
- some articles discussing this.
-
- --
- Alan Baljeu
- This signature is.
-