ir230@sdcc6.ucsd.edu (john wavrik) (07/19/90)
I must apologize to the NET for posting a personal message To: Doug Philips, Dennis Ruffer, and others Re: replies I have been occupied for several weeks with teaching an experimental course. I'm teaching every day and working on course preparation at night. This will be going on for the next month. Some of your recent postings directed specific questions at me, but were framed in a way that suggested a posting rather than email response. Please be understanding if I do not make an immediate reply. I'm not being rude -- things have just gotten backlogged. Here are some of the things that I want to post responses to and a quick outline of my reply. 1. Bill Bouma asked over a month ago about the distinction I was making about "add-on" and "built-in" features. He also raised some questions about "clashes" if add-on features are used. I wanted to clarify the distinction and to explain how features of the Forth dictionary structure (in particular that re-definitions of a word do not change the earlier version) and the vocabulary mechanism minimize these problems in Forth [as opposed to a language like LISP]. The Forth dictionary structure makes it easy to do some things in Forth that are hard to do in other languages. 2. Wil Baden posted some articles involving the generation of permutations. He concluded them with a question "Is Forth a good language for this?". I have been trying to reach him by email to clarify what he is asking. If the question is "Is Forth a good language for developing mathematical algorithms?" my answer is that it is EXCELLENT. I've done some work with permutations to illustrate my point -- but I'm waiting for some private discussion with Wil before posting them. 3. The >IN vs EVALUATE controversy provides an excellent example of what happens when you try to standardize at too high a level. A language like Forth is made weaker by adding features at a high level and removing them at a low level. I want to develop this theme. 4. The '79 and '83 Standards. I tried to post an objective analysis of the '83 Standards. I have strong objections to both of these Standards which, in my mind, call into the question whether Forth can be defined by a Standard at all. (My current view is that Forth must be defined by a model.) 5. Standards and portability. My bottom line for a Standard is: (1) Can you write significant programs in it (2) If you write a significant program in it will it run without change on other Standard systems My next to bottom line is (3) Is the Standard simple and beautiful (4) Does it preserve the essential character of Forth I'd like an explanation from those of you who think that you can take one of your current programs and will be able to run them without change on anybody's ANSI system. Neither the BASIS documents I've read, the sample implementation I've tested, the conversations I've had, my attendance at an ANSI meeting, or the postings to this newsgroup lead me to believe this. [Download the sample system from GEnie and try it yourself.] 6. Portability Forth must be the only computer language community where the concept of portability must be explained and its desirability sold. If Forth does not survive, this will be the major reason. Maybe we should run through this again -- with strong agreement from even those who disagree about everything else. 7. Postings of Bob Berkey and others re arithmetic It might be a good idea to do an article on arithmetic in the realm of mathematics to see how it compares with what goes on in a computer. 8. Questions raised in response to my posting about Forth-83 I have not had a chance to carefully read these yet. John J Wavrik jjwavrik@ucsd.edu Dept of Math C-012 Univ of Calif - San Diego La Jolla, CA 92093
peter@ficc.ferranti.com (Peter da Silva) (07/20/90)
If Forth is to survive, long term, I suspect that it will only be in the form of Postscript. This is a pity, but judging by the tower of babel it's become it's almost a mercy-killing. -- Peter da Silva. `-_-' +1 713 274 5180. 'U` <peter@ficc.ferranti.com>
jjwoehr@nyx.UUCP (jack joseph woehr) (07/23/90)
In article <YQS4EOE@xds13.ferranti.com> peter@ficc.ferranti.com (Peter da Silva) writes: >If Forth is to survive, long term, I suspect that it will only be in the >form of Postscript. This is a pity, but judging by the tower of babel it's >become it's almost a mercy-killing. >-- Oh, Pete, Pete ... what gilded nonsense! Forth is growing dramatically in its share of realtime control projects. Economically, Forth is drawing more dollars this year than ever before. It just ain't happening in that ivory tower all you desktoppers and mainframers live in :-) The control programming market in the U.S. has been estimated at $2 billion dollars annually. (EDN) Estimates of Forth's share of that market range from 9% to 15%. To match your rhetorical flight, let me quote Gary Betts (Universal Synergetics, memb. ANS X3J14) about the prospect of Forth surviving on in realtime control projects only: "We had the dinosaurs, but now they are all gone. Yet the lowly chicken remains to this day." =jax= ``onothimagain''
peter@ficc.ferranti.com (Peter da Silva) (07/25/90)
In article <1640@nyx.UUCP> jjwoehr@nyx.UUCP (jack joseph woehr) writes: > The control programming market in the U.S. has been estimated > at $2 billion dollars annually. (EDN) Estimates of Forth's share of > that market range from 9% to 15%. Hey, that's the business I'm in, and that doesn't agree with my own observations. I'd appreciate some background on where those figures are coming from. I haven't worked in Forth on SCADA for years, and haven't heard of a single Forth sales win in a new SCADA product in some time. -- Peter da Silva. `-_-' +1 713 274 5180. 'U` <peter@ficc.ferranti.com>