Re: Event based language, does it exist?

cbbrowne@acm.org (Christopher Browne)
15 Sep 2000 01:40:54 -0400

          From comp.compilers

Related articles
[15 earlier articles]
Re: Event based language, does it exist? mac@ac.valley.net (2000-09-08)
Re: Event based language, does it exist? burow@ifh.de (Burkhard Dietrich Burow) (2000-09-08)
Re: Event based language, does it exist? cfc@world.std.com (Chris F Clark) (2000-09-08)
Re: Event based language, does it exist? mwh@gradient.cis.upenn.edu (2000-09-08)
Re: Event based language, does it exist? loewis@informatik.hu-berlin.de (Martin von Loewis) (2000-09-08)
Re: Event based language, does it exist? Norman_member@newsguy.com (Norman Culver) (2000-09-08)
Re: Event based language, does it exist? cbbrowne@acm.org (2000-09-15)
Re: Event based language, does it exist? henter@wxs.nl (Peter Stevens) (2000-09-21)
Re: Event based language, does it exist? lex@cc.gatech.edu (Lex Spoon) (2000-09-21)
Re: Event based language, does it exist? trollet@skynet.be (Atle) (2000-10-08)
Re: Event based language, does it exist? rog@vitanuova.com (2000-10-10)
| List of all articles for this month |

From: cbbrowne@acm.org (Christopher Browne)
Newsgroups: comp.compilers,comp.lang.misc
Date: 15 Sep 2000 01:40:54 -0400
Organization: Giganews.Com - Premium News Outsourcing
References: 00-08-132 00-09-004 00-09-051
Keywords: design

In our last episode (8 Sep 2000 02:16:01 -0400),
the artist formerly known as Dan Cohen said:
>> In 1986, I interviewed a company in Markham Ontario that was developing
>> its own better-than-Unix computer. The hardware (!!!) was strictly
>> event-driven, and the software (at the API level) was simulated
>> event-based. I'm sorry but I forget the company name, and they're not
>> there anymore.
>>
>> I didn't get the chance to join their team, so I wasn't allowed to see
>> the language design.
>
>The event-passing protocol was:
>
>- The sender "throws" a message which carries some data
>- Any receiver may "catch" the message, which starts its procedures on
>the data
>
>I remember at the time, the boss made a 'throw' gesture, which seemed
>outrageous. None of the team made a 'catch' gesture, which seemed
>hopeless. But the system worked.


Sounds a whole lot like:


a) QNX, an RTOS that does everything
<http://support.qnx.com/support/docs/qnx4/sysarch/intro.html> via
synchronous message passing;


b) IBM MQSeries, which does everything via asynchronous message
queueing.


In both cases, the "most native" underpinnings are that the event
passing schemes represent function calls in C.


I could readily see implementing some language "structures" via
something like unto Lisp macros that would transform messaging into
something indistinguishable from other native control structures.


That would probably also work out well in Smalltalk, which also
happens to do just about everything via objects sending messages to
other objects.



Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.