Path: utzoo!attcan!uunet!lll-winken!lll-tis!helios.ee.lbl.gov!pasteur!ames!amdcad!light!bvs From: bvs@light.uucp (Bakul Shah) Newsgroups: comp.lang.c Subject: Re: the D programnming language Summary: A neat idea Message-ID: <1988May26.161419.7479@light.uucp> Date: 26 May 88 23:14:17 GMT References: <6622@cit-vax.Caltech.Edu> <4149@haddock.ISC.COM> <6206@sigi.Colorado.EDU> <4186@haddock.ISC.COM> <6293@sigi.Colorado.EDU> <7972@brl-smoke.ARPA> Reply-To: bvs@light.UUCP (Bakul Shah) Organization: Light Systems, Mountain View, CA Lines: 48 In article <7972@brl-smoke.ARPA> gwyn@brl.arpa (Doug Gwyn (VLD/VMB) ) writes: >In article <6293@sigi.Colorado.EDU> swarbric@tramp.Colorado.EDU (Frank Swarbrick) writes: >>But there's always D! > >I've been thinking a bit about what C could be like were it to be >designed today. I think it could be made noticeably smaller with >cleaner semantics (for example: strict, extensible typing; reserved >name spaces). Lots of stuff that people have been suggesting for >"D" could be left out and a better language would result. But who >is going to do this? Wirth keeps coming up with blah languages, >Ritchie has other fish to fry, etc. I'd like to try but am not in >a position to do so. > >C++ does not fit this notion of a C replacement, by the way, no >matter how useful it is. I agree with Doug totally! Privately I have been calling such a hypo- thetical language ``C-- '' (it'll have the value of C but be smaller :-) Call it what you will but the important thing is it has to be typesafe and have a small set of _simple_ and powerful _orthogonal_ features. This is a favorite subject with me. Like Doug and many others, I too have lots of ideas on what should (and should _not_) go in D which I'd like to discuss and try out in a prototype compiler except that there is only a limited amount time for playing. However, with enough people contributing to the design, specification and prototype compiler of D, may be we will have something in a year or two. If there is sufficient interest, we ought to start a mailing list as that will be a better forum for discussing the design of D. The goal would be not to collect a zillion ideas but to define the smallest possible language that meets a small set of objectives. One very desirable goal would be a simple mapping from the (more or less) complete ANSI C to D. A mapping from D to a small subset of ANSI C (or K&R C) will also be useful. Of course, any such discussion has to start with what features to take out of C, or alternatively, what features can be replaced with more general features. For example, if constructors are introduced, we can get rid of the initialization rules (even a function definition can be considered a special case of this). Now if anyone can remove pointers from the language he gets all the stars he removed _AND_ my vote to be the president of DSF (D language Specification Foundation)[1] ---- Bakul Shah <..!{ucbvax,sun}!amdcad!light!bvs> [1] Put a smiley here.