unni@sunset.sm.unisys.com (Unni Warrier) (08/01/87)
>From sdl@mitre-bedford.ARPA Sat Aug 1 13:18:36 1987 Posted-From: The MITRE Corp., Bedford, MA Subject: interfaces to simulation engines [complaining about PAWS interface] > I do yearn for a better language / interface, so that creating the > simulations would be much more easy (object-oriented, graphical interface > type of thing). Does anyone out there know of better tools that we could > all hear about? We've been developing a distributed simulation support framework in Ada. The approach is similar to Misra-Chandy. The major difference is that blocking & deadlock conditions are resolved "on demand" from the blocked node(s), by polling of their predecessor nodes. This approach (developed by Usha Chandra @ U Texas) should eliminate much overhead associated with the usual Chandy-Misra null message traffic. Anyway, we've got a reasonable interface to the simulation machinery, written out as Ada generic package specifications. You write your model's behavior as a collection of Ada tasks. These get "plugged" into the node-and-channel network (via the package spec interface), and away you go. Thus, both the interface to the framework, and the modelling components themselves, can be coded in Ada in an object-oriented style. Doing such models in Ada, however, soon makes you wish that Ada had more direct support for subclassing/inheritance (in the style of, say, Smalltalk-80). This would substantially cut down on the amount of code you need to write. (I believe there is no fundamental conflict between strong typing and inheritance; programming languages are under development that support both. Ada could be modified to support both too.) (*** flame on ***) Since the Ada language comes up for review every five years, I would like to see support gathered for improving Ada's support for inheritance by 1993. The first step is developing a proposal for how to add inheritance to Ada without breaking the language. Anyone interested? (*** flame off ***) ***** Message 2 ************ Date: Sat, 1 Aug 87 16:18:32 EDT One more thing I forgot to mention. The rigor enforced by Ada's semantic rules on the package interfaces, has enabled us to develop a systematic (though not yet automated) methodology for model construction, involving the aggregation of the modelling components together with the node-and-channel simulation framework. We believe that this methodology is sufficiently systematic to permit eventual automation of the process, using modern graphics workstations. Doing this is one of our goals for FY88. Steven Litvintchouk MITRE Corporation Burlington Road Bedford, MA 01730 (617)271-7753 ARPA: sdl@mitre-bedford.arpa UUCP: ...{cbosgd,decvax,genrad,ll-xn,philabs,security,utzoo}!linus!sdl