Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxt!houxm!whuxl!whuxlm!akgua!gatech!seismo!brl-adm!brl-smoke!gwyn From: gwyn@brl-smoke.UUCP Newsgroups: net.lang.c Subject: Re: Re: Command line arguments Message-ID: <1636@brl-smoke.ARPA> Date: Sat, 8-Mar-86 17:20:57 EST Article-I.D.: brl-smok.1636 Posted: Sat Mar 8 17:20:57 1986 Date-Received: Tue, 11-Mar-86 01:08:27 EST References: <1463@brl-smok.UUCP> <8200001@ztivax.UUCP> Reply-To: gwyn@brl.ARPA Organization: Ballistic Research Lab (BRL) Lines: 35 There have been articles recently pointing out that there is no "command line" in an icon-driven environment. I believe some of the articles have missed the point about argv (which was the original topic of discussion). According to X3J11, there are two possible conforming C environments: "hosted", which must be able to provide program arguments (NOT repeat NOT "command line" arguments) and receive exit status, and "free-standing", which need not even have a main() function, let alone program arguments. It is of course possible to implement C in an icon-oriented environment as a "free-standing" implementation. Generally, one will need special library routines to access data objects anyway in such an environment, so transportability is not a big issue in this case. Process interface characteristics will depend on the specific operating environment in a non-portable way. Another possibility is to provide a "hosted" implementation in the icon-oriented environment. In this case, there must be an automatic mapping of icon-selected entities into character strings for the argv array supplied to main(). This may or may not fit into existing icon interface schemes. All this means is that it may not be possible to advertise a conforming "hosted" implementation of C in such an environment. I see no problem with that; if the UNIX-like software tool model doesn't fit, don't use it. If you're going to write programs for an object-oriented environment, how about using an object-oriented programming language rather than C. Better yet, how about generalizing a la Alan Kay to an agent-oriented environment with an appropriate user interface. Other than for implementing the environment itself, one might then be able to do without "programming languages" altogether.