Re: First-class data types

tmb@ai.mit.edu (Thomas M. Breuel)
Mon, 9 Mar 1992 11:19:36 GMT

          From comp.compilers
Related articles
Reference to "First-Class Data Type" reid@vtopus.cs.vt.edu (1992-02-18)
Re: First-class data types norman@a.cs.okstate.edu (1992-03-05)
Re: First-class data types rockwell@socrates.umd.edu (Raul Deluth Miller-Rockwell) (1992-03-06)
Re: First-class data types pk@cs.tut.fi (1992-03-06)
Re: First-class data types kend@data.rain.com (1992-03-05)
Re: First-class data types tmb@ai.mit.edu (1992-03-09)
Re: First-class data types norman@a.cs.okstate.edu (Norman P. Graham) (1992-03-11)
| List of all articles for this month |
Newsgroups: comp.compilers
From: tmb@ai.mit.edu (Thomas M. Breuel)
Keywords: design, types
Organization: MIT Artificial Intelligence Lab
References: 92-02-085 92-03-029
Date: Mon, 9 Mar 1992 11:19:36 GMT

kend@data.rain.com's wrote:
      For example, in the C language function *pointers* are first class,
      but functions are not--you cannot create an unnamed function.


The crucial difference between Scheme and C wrt. functions is that Scheme
lets you create non-trivial closures at runtime, while C does not even
have notation to let you express non-trivial closures.


Closures are a very powerful and useful language feature. But Scheme's
terminology wrt. "first class functions" is unnecessarily "cutesy".
Functions themselves are no more first class in Scheme than they are in C.


Thomas.
--

Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.