wmb@MITCH.ENG.SUN.COM (09/07/90)
> > Chuck [Moore] objected to Unix on the grounds that he thought that he could > > do a better job of writing many or all of the Unix utilities and drivers .. > (I add in passing that the GNU project is doing just what CM suggested, > though that is more a side effect of their real goal.) One might presume that CM might object to using GNU, or indeed to using anything written by someone else, on the same grounds. Let me clarify my point: I have no objection to re-implementing a system for various economic reasons (e.g. performance, maintainability, ownership of code, portability, whatever). The thing that I meant to consider was whether to invent a new scheme or instead to adopt a proven scheme (which would probably involve some amount of re-implementation for some of the reasons previously stated.) Note that the GNU project has mostly been doing just what I suggested (re-implementing existing and familiar interfaces, not inventing new ones). As I recall, in the GNU Manifesto, Richard Stallman said something to the effect that Unix isn't the greatest thing in the world, and he could probably invent a better system, but Unix has the momentum so he's going to implement it. A compromise that succeeds has much greater value than the "greatest thing in the world, but nobody uses it". Mitch Bradley, wmb@Eng.Sun.COM
dwp@willett.pgh.pa.us (Doug Philips) (09/07/90)
In <9009070011.AA16871@ucbvax.Berkeley.EDU>, wmb@MITCH.ENG.SUN.COM writes: > One might presume that CM might object to using GNU, or indeed to using > anything written by someone else, on the same grounds. If CM will only use what CM has personally writen, its no wonder he was worried at one point about writing more than 40 programs in his life time. There is a conflict here between "doing it your way" and "doing it so that many have a chance to share in it too." In "Thinking Forth," CM suggests that arrays are an abstraction best left to the individual problem/programmer to work out. If that is true, what chance do we have with a more complicated abstract data type foundation for OOF? > Note that the GNU project has mostly been doing just what I suggested > (re-implementing existing and familiar interfaces, not inventing new ones). > As I recall, in the GNU Manifesto, Richard Stallman said something to the > effect that Unix isn't the greatest thing in the world, and he could > probably invent a better system, but Unix has the momentum so he's going > to implement it. A compromise that succeeds has much greater value than > the "greatest thing in the world, but nobody uses it". Actually I think what GNU is doing is making something better from the inside out. Many of the GNU tools go beyound the things they are replacing. Whether that is an explicit goal, a hidden agenda, or just plain good fortune I don't know. Given the relative number of systems/ programmers/users of Unix versus Forth, its easy to see why compatability would be a very big concern for GNU. The other concern that makes the analogy somewhat flawed is that GNU is not just about replacing Unix but about freeing intellectual access to software so that any improvements can be shared by all. In ten more years will GNU "Unix" really look like Unix anymore? (Semi-rhetorical question). -Doug --- Preferred: ( dwp@willett.pgh.pa.us OR ...!{sei,pitt}!willett!dwp ) Daily: ...!{uunet,nfsun}!willett!dwp [last resort: dwp@vega.fac.cs.cmu.edu]