Xref: utzoo comp.sys.next:6869 comp.software-eng:3863 Path: utzoo!attcan!uunet!cs.utexas.edu!rutgers!mephisto!psuvax1!xavier!gessel From: gessel@cs.swarthmore.edu (Daniel Mark Gessel) Newsgroups: comp.sys.next,comp.software-eng Subject: Re: Software-IC Catalog Keywords: Objects Standards Publication Objective-C Message-ID: <1W7LTWR@xavier.swarthmore.edu> Date: 4 Jul 90 20:35:02 GMT References: <5320@stpstn.UUCP> Sender: news@xavier.swarthmore.edu (Usenet News) Organization: Swarthmore College, PA, USA Lines: 58 In <5320@stpstn.UUCP> cox@stpstn.UUCP (Brad Cox) writes: >In article gessel@cs.swarthmore.edu (Daniel Mark Gessel) writes: >>Is the NeXT community ready for a publication which would list specifications >>of objects for sale (or for free), and possibly have discussions of some >>standard interfaces for kinds of objects on top of NeXT Step? >That I would encourage such thought should be obvious to all concerned. But >I'd encourage you to think about pushing it much further than an ordinary >paper or electronic catalog. >Paper catalogs are sufficient for tangible components, like silicon chips >and plumbing supplies. But for intangible components like Software-ICs(TM) >I think that we must apply some technology to make intangibles observable. >A simple way of doing this (which soon leads to fascinating implications >regarding how the semantics of natural languages are established, with >direct implications to specification/testing languages) is to also publish, >as an precondition to the catalog, a library of executable gauges (test >procedures) that detect whether a Software-IC that purports to comply with >some abstract specification, does in fact do so within some tolerance for >non-compliance. . . . >Brad Cox; cox@stepstone.com; CI$ 71230,647; 203 426 1875 >The Stepstone Corporation; 75 Glen Road; Sandy Hook CT 06482 Do you suggest that this library should be defined along with the standard interface, and published along with (either in paper or electronically, or both)? Or, should the "catalog" be a set of programs which actually exhibits the compliance? That is, I don't see why anything more than an oridinary paper or electronic catalog is required. What you are suggesting, it seems to me, is that specifications should also require some method of "proof" of compliance along with the specification itself. I think this is a great idea, and I think it should be incorporated into any good specification. It has been suggested that the NeXT User's Journal might be a good place to publish specifications, although I would expect that they would need to be developed by other organizations. Possibly a newsgroup or mailing list. Standards will come about, I expect, as interfaces are defined for commercial libraries that are marketed. There is the possibility (and danger of bad) de facto standards, dependent on libraries which make it out early and are used widely, although copyright issues may make for problems there (another reason to create standards fairly early in the game). Standards will be developed, by one means or another, but I think taking the situation in hand directly will help to make standard interfaces better from the beginning. Dan -- Internet: gessel@cs.swarthmore.edu UUCP: {bpa,cbmvax}!swatsun!gessel