home *** CD-ROM | disk | FTP | other *** search
- Path: ornews.intel.com!news
- From: Patrick Logan <patrick_d_logan@ccm.hf.intel.com>
- Newsgroups: comp.lang.java,comp.lang.c++,comp.lang.smalltalk
- Subject: Type systems, Dylan, Smalltalk, etc. (Was: Re: Will Java...)
- Date: Thu, 18 Apr 1996 13:24:32 -0700
- Organization: Intel
- Message-ID: <3176A500.4CAA@ccm.hf.intel.com>
- References: <31682FFE.2781E494@bbn.com> <dbell-1 <3171810F.2E2@funsys.se> <4l0f6o$sec@nkosi.well.com> <Dpyro4.8o9@news.hawaii.edu>
- NNTP-Posting-Host: jlchesne.intel.com
- Mime-Version: 1.0
- Content-Type: text/plain; charset=us-ascii
- Content-Transfer-Encoding: 7bit
- X-Mailer: Mozilla 2.01 (WinNT; I)
-
- Peter Hinely wrote:
- >
- > This type system allows Dylan programs to be prototyped rapidly without
- > type information. Type information can later be added in, which the
- > compiler can use to optimize the program better, giving you the best of
- > both worlds: rapid development and good performance in your final program.
-
- The Dylan type system is a *mix* of both worlds. It has yet to be
- proven that it actually is the *best* of both worlds.
-
- There are systems out there that do comprehensive type inferencing to
- the point they appear to be more than competitive with C. (After all
- we still have to compare everything to C, don't we?)
-
- Consider the Stalin system for Scheme from U. Toronto.
-
- This is approaching more of what I would consider the *best* of both
- worlds: getting the performance and security of a statically
- typed language without ever having to explicitly name a type.
-
- Extrapolating the Stalin experience (now there is a scary term)
- to Smalltalk would be a good research topic, IMO.
-
- --
- Patrick_D_Logan@ccm.hf.intel.com
-
- "OO is a tool, not a religion, and not a philosophy."
- -Robert C. Martin
-