home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.realtime
- Path: sparky!uunet!haven.umd.edu!darwin.sura.net!zaphod.mps.ohio-state.edu!cs.utexas.edu!uwm.edu!spool.mu.edu!agate!ames!sun-barr!sh.wide!wnoc-tyo-news!news.u-tokyo.ac.jp!ccut!s.u-tokyo!is.s.u-tokyo!jeff
- From: jeff@is.s.u-tokyo.ac.jp (Jeff McAffer)
- Subject: Re: novel idea?
- References: <erempel.726720401@sol.UVic.CA>
- <1993Jan11.195020.25839@frog.CRDS.COM>
- Nntp-Posting-Host: water
- Message-ID: <1993Jan12.130002.13110@kei.is.s.u-tokyo.ac.jp>
- Reply-To: jeff@is.s.u-tokyo.ac.jp
- Organization: University of Tokyo / Object Technology International
- In-Reply-To: miket@frog.CRDS.COM's message of 11 Jan 93 19:50:20 GMT
- Sender: news@kei.is.s.u-tokyo.ac.jp (Usenet News System)
- Date: Tue, 12 Jan 1993 12:59:58 GMT
- X-Bytes: 2045
- Lines: 43
-
- In article <1993Jan11.195020.25839@frog.CRDS.COM> miket@frog.CRDS.COM (Michael C Tiernan) writes:
-
- > You *MIGHT* want to consider the use of a psudo code at this
- > point, THEN, you could switch systems for a process. However,
- > you'd then need an interpreter running on each machine which is
- > identical from the appliction's point of view. And a zillion
- > other complexites build up from there. ^^^^^^^
-
- That is not quite true. For example, I have used embedded Smalltalk
- on VMEbus machines for realtime military signal processing (Radar)
- applications. Yes, there are some issues. It happened that the
- machine was homogeneous shared memory but it would also work on groups
- of heterogeneous distributed memory machines. I would not have wanted
- to approach that project in any other language/environment.
-
- For those who don't know what Smalltalk is: It is fully
- Object-Oriented language/environment that *typically* is compiled to
- bytecodes (similar to p-code) which is then run by a virtual machine
- (interpreter). many of the new Smalltalk compilers actually compile
- directly to machine-code. BTW, if you buy one of Tektronix new
- digital osciliscopes you are really buying a 680?0 running a fully
- interpreted Smalltalk! Who says you can't do real-time stuff with
- interpreted languages?
-
- Anyway, I am not trying to sell Smalltalk. Just point out that by
- looking at things a little higher level or from a different
- perspective many of these "problems" go away.
-
- I would argue that if you are caring about machine instructions there
- is something wrong with your design. Especially if you have multiple
- processors.
-
- BTW, for our signal processing app, we could actually develop code on
- a 486 and move it to the target (040) with no changes. (well, ok,
- little tiny changes because on the 486 we had to model our custom
- hardware). Doesn't take long to figure out the economics of 15
- engineers, $100,000 testbed and $4,000 486s.
-
- But I digress
-
- --
- ato de, |m -- spice is the irony of life! thks mm
-