Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!uakari.primate.wisc.edu!crdgw1!uunet!bywater!scifi!watson!arnor!andreadoria!strom From: strom@watson.ibm.com (Rob Strom) Newsgroups: comp.lang.misc Subject: Re: A Hard Problem for Static Type Systems Message-ID: <1991May3.154615.28787@watson.ibm.com> Date: 3 May 91 15:46:15 GMT References: <566@eiffel.UUCP> <2672@optima.cs.arizona.edu> Sender: news@watson.ibm.com (NNTP News Poster) Reply-To: strom@andreadoria.watson.ibm.com (Rob Strom) Organization: IBM T.J. Watson Research Center Lines: 23 Nntp-Posting-Host: andreadoria In article , olson@juliet.ll.mit.edu ( Steve Olson) writes: |> |> Um, well, in my opinion, the extra keystrokes involved are the weakest of |> the arguments against static typing. I mean, the botom line isn't keystrokes, |> its total programmer effort. And of course, in many applications, machine |> effort must be considered also. I don't mind typing the declarations so much |> as I mind being forced to specify my data objects at the bits 'n' bytes level. |> ("32 or 64 bits, pal, thats all we got here, and you better not get 'em |> mixed up either!") |> The above isn't an argument against static typing. It's an argument against type systems which fail to be abstract and which therefore make representation visible as part of the type information. Static type systems lacking that property are immune from this criticism. (Insert subtle plug for our Hermes language here :-) A dynamic type system which gave run-time type errors when users tried to add 32 and 64 bit numbers together would be just as bad. -- Rob Strom, strom@ibm.com, (914) 784-7641 IBM Research, 30 Saw Mill River Road, P.O. Box 704, Yorktown Heights, NY 10958