benson@odi.com (Benson I. Margulies) (10/26/89)
On consideration, I see little purpose in wasting bandwidth in point-by-point rebuttal of some of the replies to my latest item. However, there are some clarifications I might as well post: 1) My position is a reaction to the content of the GNU manifesto and the CopyLeft. I'm not interested in arguing about what the FSF does, and thus about operational defininitions. I'm interested in it and RMS's publicly stated political and economic goals. I can easily imagine someone dismissing much of that material as unlikely to happen and this irrelevant, but I currently choose not to. So I for one am not interested in statements of the form "The purpose of the FSF is to create GNU," or "I think freely usuable software is a good thing, and the FSF is producing it." I don't claim that those are stupid attitudes, just not relevant to the discussion at hand. 2) My comments about the "real world" weren't intended to insult anyone. Rather, to compare Lotus 1,2,3 to and Unix product. I have friends at Lotus. That product is as bug free as it is, and it is plenty bug free, in part because of an enormous investment in testing, design-for-test, and plain old QA. A typical spreadsheet user would find source utterly unhelpful in a buggy product. The overall software economy moves towards software for non-programers. Someone raised the point, "but all that QA is just a service. You could make a business in the FSF world of fixing all those bugs." a) well, but not for long. Once the bugs were fixed, all your competition would have all of your work for free. b) at the risk of agreeing with Grandi, you don't slap testing on afterwards. You make an investment in building it in from the start. 3) "Unix vendors don't do shit for QA" Initially, the Unix marketplace has been a programmer/hacker marketplace. It tolerated shitty code. No one seems to have ever made a decision to resist Sun's prices in order to buy someone else's workstation with a non-buggy Unix. Why? Perhaps because Unix from AT&T and BSD was so incredibly buggy that no vendor was about to rewrite large parts of it to squeeze out the bugs. If they did, they would bust too many application that depend on some of the bugs. So Unix is a terrible model of why intellectual property rights are needed to produce high quality software. I tend to suspect that if it weren't for Unix, there wouldn't be an FSF. The buggy state of Unix leads to great frustrations in programming users who don't have the source. The source is difficult to get, because AT&T's $60K charge prevents even friendly vendors from making it available. I once worked on Multics. Multics came with complete source. Copyrighted. If you didn't want to wait for us to fix a bug or add a feature, you could do it yourself. Traditionally, Evil Proprietary OS vendors either gave you the source or at least charged a plausible fee for the privilege. Only the Wonderful Open Unix system brought us a minimum fee of $60K. -- Benson I. Margulies
lee@uhccux.uhcc.hawaii.edu (Greg Lee) (10/27/89)
From article <1989Oct26.133749.15831@odi.com>, by benson@odi.com (Benson I. Margulies): " ... " Someone raised the point, "but all that QA is just a service. You " could make a business in the FSF world of fixing all those bugs." " " a) well, but not for long. Once the bugs were fixed, all your " competition would have all of your work for free. ... Really? Why not copyright your bug fixes or enhancements? I can see that your work might have to carry the copyleft and might have to be distributed in source form. I don't know why that would prevent you from also copyrighting it yourself -- it's your work. Greg, lee@uhccux.uhcc.hawaii.edu