ir230@sdcc6.ucsd.edu (john wavrik) (07/03/90)
Doug Philips writes: > Personally I'd prefer a toolshop that already has some standardized > tools in it. Not mandatory, but available. Presumably there is > enough experience with building houses that I could reasonably expect > to find a hammer in the toolshop already. Even if I have to make my > own sledge for pounding in the stakes for the concrete forms, I'd > expect to find a utility hammer at least already there. I should be > able to remove it as well. I think the idea of "extended wordsets" > fits this bill perfectly. We all have visions of what Forth should have become. Mine was that the foundations would have stabilized about 8 years ago to the point where one could write significant applications which would run on any Forth. This would have led to the development of a refereed library of high quality tested tools that one could buy at modest cost (but which could be freely used in applications). This, in turn, would have led to the concept of "Tools You Can Modify" -- tools explicitly written to facilitate adaptation. The net result would have been a programming environment both more flexible and more powerful than anything conventional languages can provide. I think that what has actually happening is something like this: Most toolmakers have been using hexagonal heads on their nuts and bolts -- but one major vendor uses pentagonal heads. Now comes the Standards meetings. Each of them have made something called a "box wrench", but obviously it has a 5 sided opening for one guy and a 6 sided opening for everyone else. The six sided people point out that their design has the advantage of allowing open-end wrenches -- but to no avail (the 5 sided guy believes that he can tighten a nut 3% faster if it has a 5-sided head -- and he threatens to walk out unless 5-sided nuts are made Standard). After hours of deliberation, they decided to leave the issue of how many sides a nut has "implementation dependent". But one among them says that it would be inappropriate to do this with something so basic as nuts and bolts. Finally someone comes up with a brilliant compromise: if a wrench has a 30 sided opening it will fit over both types of nuts. [No matter that no one has ever made such a thing -- and if they did they'd find it would not support much torque and has a tendency to round off the corners of nuts] Finally, there was the matter of a name -- both groups had been using the name "wrench" -- so they decided to call the new object a neutral name: "sponge". And what about the useful open-end wrenches? "They are obviously non-Standard since they do not conform to existing practice, which includes both 6 and 5-sided nuts. Now of course, anyone who knows what is involved in getting this toolshop idea off the ground knows that after 2 years of deliberation we should not only have agreement on the shape of nuts, but on thread sizes and diameters for bolts and hundreds of other basic things. Instead we have a 30-sided sponge and no open- end wrenches. (But don't worry, the sponge is in the extended wordset, so you don't have to use it if you don't want to) John J Wavrik jjwavrik@ucsd.edu Dept of Math C-012 Univ of Calif - San Diego La Jolla, CA 92093
dwp@willett.UUCP (Doug Philips) (07/04/90)
ir230@sdcc6.ucsd.edu (john wavrik), in <11713@sdcc6.ucsd.edu> writes: > We all have visions of what Forth should have become. Mine was that > the foundations would have stabilized about 8 years ago to the point > where one could write significant applications which would run on any > Forth. Meaning we should have stuck with Forth-83, Forth-79, Fig-Forth, or perhaps done the ANSI effort earlier? I think you have a nice idea here. I curious as to why you think it didn't happen? Was it even realisticly possible? > This would have led to the development of a refereed library > of high quality tested tools that one could buy at modest cost (but > which could be freely used in applications). This, in turn, would have > led to the concept of "Tools You Can Modify" -- tools explicitly > written to facilitate adaptation. Aha, perhaps what we need here is GNU-Forth? Free, widely distributable, add on libraries? > The net result would have been a > programming environment both more flexible and more powerful than > anything conventional languages can provide. I'm curious about your "Tools You Can Modify" concept. It sounds to me like Object-Oriented Programming. Is that it? If not, how do you see it as different from OOP? ... > Instead we have a 30-sided sponge and no open- > end wrenches. (But don't worry, the sponge is in the extended > wordset, so you don't have to use it if you don't want to) I think Mitch's rebuttal says everything I have to say about this from a technical point of view. What interests me now is how this sort of "political" divisiveness ends up happening. Is it really the case that there are only 16% 5-sided bolts or is it 50% as Mitch suggests. I have seen NO evidence to support those numbers one way or another. I still haven't seen any answer from anyone on the TC about how they go about quantifying that kind of information. I am neither more inclined to believe John nor Mitch at this point, since both are merely making assertions to support their positions. Doesn't ANSI spell out in any procedural detail how the TC is supposed to assess "existing/common practice?" Does it instead require the TC to submit a proposal for how it will do that? Does the TC just get to say what it wants to on that topic, unchallenged? I find it hard to believe that the various sides of this issue are merely promoting their own self interest by lying with statistics, but don't have the information to draw any other conclusions. If this kind of bickering has been going on since the beginning of the "standard" effort, I can see why many of the TC committee would be fed up. I also don't understand why they wouldn't have adopted some kind of formal policy about what was to qualify as common practice and just cut past all the BS. Is it because ANSI will let anyone who wants to be "sit" on the committee and just nay-say anything that comes by, if they feel like it? Are John and Gary really saying that the ANSI process waters down standards so much that they are worthless? -Doug --- Preferred: willett!dwp@hobbes.cert.sei.cmu.edu OR ...!sei!willett!dwp Daily: ...!{uunet,nfsun}!willett!dwp [in a pinch: dwp@vega.fac.cs.cmu.edu]