Mitch.Bradley@ENG.SUN.COM (05/02/91)
> >P.S. I would also note that it might prevent some of the useful compromises > > that have taken place, i.e. WORDLIST and floored vs. symmetric division. > In your example, WORDLIST is the only thing that wouldn't have made it- > floored and symmetrical division has been around and implemented ... The division example is irrelevant; it isn't a new word, so it wouldn't have been subject to the proposed rules. Some other things that wouldn't have made it: files, most of the floating point wordset, error handling, memory allocation. The criteria that the committee used for common practice was less restrictive and less precise: If a lot of systems have chosen to solve a particular problem, and committee members agree that it is an important problem, the it is fair game to synthesize a wordset based on knowledge of the different solutions that have been used in the past. Brad's proposed criteria requires exact semantic match among major systems. This hardly ever happens in the Forth world, except for words that are already standardized (and sometimes not even then!). Mitch.Bradley@Eng.Sun.COM