Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.2 9/18/84; site dataioDataio.UUCP Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxt!houxm!vax135!cornell!uw-beaver!uw-june!entropy!dataio!bright From: bright@dataioDataio.UUCP (Walter Bright) Newsgroups: net.micro Subject: Re: Binary Compatibility 80286 Message-ID: <844@dataioDataio.UUCP> Date: Fri, 25-Oct-85 13:02:56 EDT Article-I.D.: dataioDa.844 Posted: Fri Oct 25 13:02:56 1985 Date-Received: Sat, 26-Oct-85 08:13:53 EDT Reply-To: bright@dataio.UUCP (Walter Bright Organization: Data I/O Corp., Redmond WA Lines: 39 References: In article <2380@brl-tgr.ARPA> gwyn@brl-tgr.ARPA (Doug Gwyn) writes: >[Software houses] merely need to compile their sources into the appropriate >binary for the target machine and ship [it]. This need not even imply >running the particular target systems; cross-targeted software generation >systems can be used. I routinely compile code for my DMD on a host >with a totally different architecture. Yah, but are you SURE that the binaries you ship work if you never ran them? >There is no more reason to insist on binary standards for the mass market >than there is to insist that all ball-point pen refills be interchangeable. Yes there is. Reasoning is: 1) Small software developers can't afford to buy n machines just so they can compile their product on them. Especially when they cost $10000 plus. Borrowing machines doesn't work, as you can't support software when you run out of favors borrowing the machine. 2) Large software developers have a problem with compiling the product on n machines. This is inventory. One has to make multiple releases, and stock multiple versions, etc. Large inventories are the enemy of profits. Many companies also have large expenses associated with releasing a product. 3) Compiling on another machine implies testing on that machine. For interactive programs, testing cannot be automated and thus becomes a large burden for each new binary produced. 4) Same arguments apply to multiple distribution medias. 5) The success of the industrial revolution depended on standardization and production of millions of identical parts. Where would we be if every manufacturer designed their own bolts and nuts? They all accept design compromises to use standard parts. I suggest that the software industry must make the same kinds of compromises to survive.