koopman@a.gp.cs.cmu.edu (Philip Koopman) (01/21/91)
In article <15788@sdcc6.ucsd.edu>, ir230@sdcc6.ucsd.edu (john wavrik) writes: > [[entertaining stuff about dating deleted]] > The idea of dating sounds like a good idea. I wish that there > were a way of making a formal proposal to the ANSI team to allow > a dating period before they declare this phase of their work finished. > (Someone should make a test implementation -- they were doing that > for a while -- and let the Forth Community see what would be involved > if she were to become a companion for life.) Maybe I am wrong, but here is what I am assuming will happen: The "proposed standard" (dpANS? I don't remember the acronym) will be released for review by the community. Farsighted vendors who want to ensure they have a future will immediately release Forth systems that conform (it should be easy, since the standard is probably changing less with each revision now). Then, people in the Forth community who are worried about being "stuck" will use those compilers for their next big piece of code. If it doesn't work, they'll scream loud and long. If it does work, they have a leg up on using the new standard Forth for all their future projects. Presumably, there will be enough time for this process to happen (a few months). Those who don't make an effort to actually use the new (proposed) standard for real work will not have earned the right to complain when the final version is issued. Those who do use it will be able to contribute to a final product based on real-world experience, to the benefit of all. Phil Koopman koopman@greyhound.ece.cmu.edu Arpanet 2525A Wexford Run Rd. Wexford, PA 15090 *** this space for rent ***
ir230@sdcc6.ucsd.edu (john wavrik) (01/22/91)
Phil Koopman writes, > Maybe I am wrong, but here is what I am assuming will happen: > The "proposed standard" (dpANS? I don't remember the acronym) > will be released for review by the community. Farsighted > vendors who want to ensure they have a future will immediately > release Forth systems that conform (it should be easy, since > the standard is probably changing less with each revision now). > Then, people in the Forth community who are worried about being > "stuck" will use those compilers for their next big piece of > code. If it doesn't work, they'll scream loud and long. > If it does work, they have a leg up on using the new standard > Forth for all their future projects. No, Phil, you're not wrong UNDER NORMAL CIRCUMSTANCES. The situation here is very like what happens when you commission someone to build a house for you. There is a critical point when the builder declares himself done. At that point, you must agree that, in principle, the builder has accomplished what he has been asked to do. After that, you can ask him to retouch some paint -- but if the builder hasn't put in a bathroom, you must reject the entire house. Under normal circumstances the builder does something expected and produces something you have every right to expect is what you want. Suppose, however, you have a builder who periodically reports that he has had to change the design because of conflicts between the plumber and carpenter? Suppose, on one of your close inspections, you found that the hot and cold water taps were reversed -- and that there were no electrical outlets in the kitchen? The builder has not done what he was expected to do, by his own admission. Whether he has, in principle, accomplished what he was asked to do is unclear (he has, in fact, produced a different house). You have one of three choices: 1. Reject the house because it is being built in a manner that is unlikely to produce satisfactory results. 2. Try to be charitable to the builder -- and ask that he allow a more extended inspection period than usual (and be willing to fix more deep-seated defects). 3. Blindly accept the house entirely on the builder's assurance. My personal opinion: I think that Forth needs a good Standard but that the worst thing that can happen to Forth is yet another devisive Standard. P.S. I can tell you right now: the house has no bathrooms! John J Wavrik jjwavrik@ucsd.edu Dept of Math C-012 Univ of Calif - San Diego La Jolla, CA 92093