Xref: utzoo comp.sources.d:2103 news.admin:2158 Path: utzoo!attcan!uunet!seismo!sundc!pitstop!sun!decwrl!ucbvax!unisoft!hoptoad!gnu From: gnu@hoptoad.uucp (John Gilmore) Newsgroups: comp.sources.d,news.admin Subject: binaries on the net Message-ID: <4577@hoptoad.uucp> Date: 13 May 88 08:38:45 GMT References: <1574@looking.UUCP> <22099@bu-cs.BU.EDU> <392@pan.UUCP> <508@splut.UUCP> Organization: Grasshopper Group in San Francisco Lines: 32 jay@splut.UUCP (Jay Maynard) wrote: > People who argue for source-only distributions don't believe in letting > an author do as he likes with HIS work. This person who argues for source-only distributions has no trouble with letting an author do as he likes with his work -- as long as he doesn't spend other peoples' money doing it. My argument against binaries runs along one line: Binaries are very hard to maintain and very hard to port. What good is the last 30 years of progress in software engineering if we never get the sources? Maintenance is always the largest cost of any piece of software. Binaries are just throwaways; change CPU or OS and toss 'em. Don't tell me an IBM PC binary good today will be good forever...I know better. How many binary programs from PC-DOS 1.0 (before the hierarchical file system) still run unchanged? There is 8088 code that doesn't run on the 386 -- Minix is one (the 386 broke the LOCK prefix on some instructions). I don't think the net should be a medium for "free" (==> other people pay for it, not you) distribution of binary software. There are plenty of IBM PC user groups, BBS's, Fidos, and companies that charge $3/floppy for PD software, doing that. It should be a medium for the continuing evolution of software (and programmers). Binaries don't evolve; when they stop working, they go extinct. Many of the people on the net are on the leading edge of software development. Their binaries won't teach us anything, but their sources will. -- John Gilmore {sun,pacbell,uunet,pyramid,ihnp4}!hoptoad!gnu gnu@toad.com "Use the Source, Luke...."