home *** CD-ROM | disk | FTP | other *** search
- Path: ix.netcom.com!news
- From: Mike Girou <girou@parashift.com>
- Newsgroups: comp.lang.c++
- Subject: Re: Will JAVA kill C++?
- Date: Tue, 19 Mar 1996 11:16:39 -0600
- Organization: Paradigm Shift, Inc.
- Message-ID: <314EEBF7.67BE@parashift.com>
- References: <313E44EA.14D110C0@netcom.com> <4hp18v$3di@frodo.smartlink.net> <4ht8k1$t7l@epx.cis.umn.edu> <3146278D.7703E9CC@netcom.com> <314EB074.575@esec.ch>
- NNTP-Posting-Host: ix-dfw16-18.ix.netcom.com
- Mime-Version: 1.0
- Content-Type: text/plain; charset=us-ascii
- Content-Transfer-Encoding: 7bit
- X-NETCOM-Date: Tue Mar 19 11:20:28 AM CST 1996
- X-Mailer: Mozilla 2.0 (Win95; I)
-
- Oliver Plohmann wrote:
- > What I would like to figure out is whether the "method-level" strong
- > typing
- > in Java helps in reducing runtime errors compared to dynamic typing in
- > Smalltalk.
- >
-
- Good question, sure to be controversial. Bill Berg, Marshall Cline
- and I published a paper in October 1995's Communications of the ACM
- describing the development of OS/400 in C++, and one of the strong
- feelings by the developers was that strong type checking was a major
- contributor to the project's success.
-
- On the other hand, Tom Morgan and some of the people at Brooklyn
- Union Gas kept records for quite some time on the cause of various
- programming errors in their Smalltalk development. Surprisingly (?)
- enough, problems due to run-time type checking were way down on the
- list. There is anecdotal support for this finding from other ST shops
- who have kept track of where their errors come from.
-
- One version or another of these arguments have been around for many
- years, and the answer is unclear. I suggest that for large,
- monolithic systems such as operating systems, compile-time type
- checking is a clear winner, but for smaller piecemeal applications
- the payoff isn't that large. The issue doesn't seem to be how many
- KLOC of code you have, but rather how tightly integrated/interrelated
- the system is.
-
- Mike
-
- the Pascal v
- --
- Mike Girou girou@parashift.com
-