karl@sugar.UUCP (Karl Lehenbauer) (03/08/88)
In article <1002@sbcs.sunysb.edu>, vrshv306@sbcs.sunysb.edu (semyon varshavchik) writes: > If it's possible to have graphics that have correct colors on all Amys, why > can't you, oh mighty programmers, do it all the time? Amy is so open-archi- > techtured, does it really pay off to save 16 bytes of code at an expense of > having your product completely unusable on future Amys? Apart from the condescending verbiage, the guy has something of a point. (but if he wears a hat no one will notice :-) Much of the stuff that people are hacking out could and should instead be written to be well behaved. A lot of people say that they absolutely have to hack their programs out for performance; I think it's mostly because they're lazy - it's hard to write well behaved code. Here's the thing that's been bugging me, that started all this: You don't know if your program is really upward compatible unless it's been shown to work on the new or expanded machine or the new version of the operating system, etc. For example, when there were only 512K Amigas, even people who had been trying to write well behaved software didn't know if they'd gotten it right until it was run on larger memory machines. One missing MEMF_CHIP can ruin all one's compatibility efforts - and there's no way to tell until the upgrade is available. (less of a problem now than then, but still a problem) I did read recently that Apple has been testing various third party software to see how well behaved it is. The results, as I recall, have been pretty dismal - something like 10% follow all the rules, 80% are close but would require some changes and the last 10% would require massive rewrites. What I'm thinking here is that maybe Commodore could do some similar thing, giving the "Commodore well behaved seal of approval" to programs that had passed some kind of test suite. I don't know all the issues that would have to be addressed by the tests. I'm sure there are several requirements that would be very tricky to test for. I also realize that this is probably pie-in-the-sky because they (C-A) don't have the time, money or inclination to do it anyway. The point is that I would like to 1) know that my program that has been written to be well behaved is in fact well behaved and 2) encourage authors to write (by getting users to insist on) well behaved programs. -k -- "Lack of skill dictates economy of style." - Joey Ramone ..!uunet!nuchat!sugar!karl, Unix BBS (713) 438-5018