home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!gatech!paladin.american.edu!howland.reston.ans.net!spool.mu.edu!agate!anarres.CS.Berkeley.EDU!bh
- From: bh@anarres.CS.Berkeley.EDU (Brian Harvey)
- Newsgroups: comp.lang.scheme
- Subject: Re: Unspecified values in R4RS
- Date: 22 Jan 1993 18:40:46 GMT
- Organization: University of California, Berkeley
- Lines: 34
- Message-ID: <1jpf3e$2m1@agate.berkeley.edu>
- References: <BRENT.93Jan22085102@rcx1.ssd.csd.harris.com> <1jp5n7$nm@agate.berkeley.edu> <OZ.93Jan22114638@ursa.sis.yorku.ca>
- NNTP-Posting-Host: anarres.cs.berkeley.edu
-
- oz@ursa.sis.yorku.ca (Ozan Yigit) writes:
- >Brian Harvey writes:
- >
- > Leaving something unspecified in the standard invites some implementor to
- > give it a "useful" value; and that encourages users of that implementation
- > to use it; and that makes programs non-portable.
- >
- >Are there such implementations? Isn't there some that return
- >#<unspecified>, which is not a "useful" value? Can we point to real
- >implementations, instead of talking about a hypothetical one?
-
- I think this response doesn't take into account the context of my quoted
- remark. I was arguing that, e.g., DISPLAY should be required to return
- a nonprinting value, instead of "unspecified" as the standard now says.
- Right now if you write a program like
-
- (define (praise thing)
- (display thing)
- (display " is great!") )
-
- you get interactions like
-
- > (praise "Ice cream")
- Ice cream is great!#f
-
- which is not what you wanted. Then *someone else* suggested that it's better
- for the standard to leave this flexible, in case someone has a good idea for
- a useful return value. You should be asking that person (I forget who, sorry)
- to point to a real example. The real examples that *I* know are mostly about
- returning a useless but annoying value, such as #f.
-
- I suppose someone must think it useful that many implementations of DEFINE
- return the symbol being defined; that's a real example for you, but *I* don't
- find it useful, and I'd be happy to see DEFINE return the nonprinting value too.
-