From: Michael Almquist Subject: Re: SCI: Object behavior in VR Date: Fri, 09 Oct 92 10:56:03 -0700 Just some thoughts and some random babbling: >If I'm not mistaken (and someone will surely correct me if I am) >Scheme uses garbage collection to return unused cons cells to the free >list, just like LISP. Seems to me that it's unacceptable to have the >virtual world freeze for a moment to let the interpreter clean up >after itself. Yes. What can one do? SmallTalk is the only package (to the best of my knowledge) that uses generation scavanging (NOT "generational" GC - its GS). >My own choice for an interpreted language would be Forth, which has >all the advantages of LISP-like languages, plus a few more: This idea has some interesting technical issues and merit. BUT the religious battle of Lisp, Scheme, Forth, modula, Language A, Language Z could and would rage on forever. Implement it and get us some RAW numbers, then we can chat with some sort of validity. >Data-flow diagrams, such as those used by VPL's Body Electric, are >easily translated into Forth code. With a few restrictions, Forth >code can also be translated into data-flow diagrams. Yep, another good point and hopefully a good question: So? Who came down from the mountain top and said data-flow is the key and the way? I hope it isn't - like most everything, dataflow has its strengths and weakness - we need to look further. >Now the question is, why use Forth? I don't really see any advantages >myself, except those that rely on the simplicity of the interperter. BUZZ! Its obvious this statement was made without peering into the depths of Forth. Forth is a language IMHO that never did receive the praise it deserved. This religious battle is unnecessary. Its unfortunate that due to time constriants (ie. get those flashy demos up and running pronto) we cannot venture down other less explored avenues. We are still in the infantile stage. There is LOTs to yet explore and implement. Also, I eventually don't want to interface with my system via a computer language but human speech and motional cues. This is the goal to shoot for - not a BEAUTIFUL killer Forth shell. - mike PS: For those of you that are intrested in the religious debate and concepts of interpreters find and read "Programming Languages: An Interpreter-Base Approach by Samuel H. Kamin" AND/OR bop over to uihub.cs.uiuc.edu:/uiuc/kamin.distr and snag the public domain code for various free small interpreters. For those of you interested in the best garbage colletion schemes concepts etc. find Dr. Dobb's Journal, May 1990, pg. 16 "Generation Scavenging". Excellant article with excellant reference list. A MUST READ. Would others out there like to know more about "the squish" classic papers archive? I've been building/reading a VAST library of papers/books/code/etc applicable to VR/systems/etc.