neubauer@bsu-cs.UUCP (Paul Neubauer) (05/23/87)
In article <168700002@uiucdcsb>, chen@uiucdcsb.cs.uiuc.edu writes: > I received the following responses on my earlier query concerning > AT compatibles that are able to run either Gold Hill 286 or Hummingboard. [several responses] > For your comaprison, the following is a list (edited) > of the AT compatibles > that have been tested by Gold Hill (from Gold Hill's Compatibility List, > dated March 4, 1987). If you find some of the entries cryptic, > your guess on their meanings is as good as mine. > Computer 286 Hummingboard > ---------------------------------------------------------------- > Compaq Deskpro 286 yes yes > Compaq Portable II yes yes > Compaq Portable III ? ? > Compaq 286 Portable yes yes > Kaypro 286I *** ? > PC Limited 286 no ? > Zenith 241 BIOS V1.7+ ? > Zenith 248 BIOS V1.7 only ? This is an extremely interesting and strange list. Apparently when Gold Hill says that their Lisp works on IBM's and "100% compatibles" they mean 100% and not 99%. I had been thinking very seriously about buying a low-priced but high-speed AT clone and GC286, but I certainly have to worry about what these results mean. Some of the systems on which GCL does not work, or works with some restrictions, are very highly thought of for compatibility, e.g. the Kaypro and the Zenith. What I had also gathered was that the AT clones in general tended to be reasonably compatible, even at the low end. Interestingly, too, yesterday I received a solicitation from Microsoft to buy their OS2 developer's package, which came with a list of some suggested development systems. Among those suggestions from Microsoft were the two Zenith models above, but only with BIOS versions 1.9 and later. In other words, it looks like Gold Hill's possible systems and those that are to be supported by Microsoft are mutually exclusive. This does not inspire confidence on my part. If Gold Hill's products are this hardware dependent and poorly behaved, I have to wonder what Gold Hill is planning to do about the new PS2's. Will they create their later versions to run only on them? I realize that buying GCL does not give one a Lisp Machine, but ergonometrically speaking, I for one am certainly not willing to even consider using EMACS (or GMACS) with a keyboard that has NEITHER the CTRL key nor the ESC key in any reasonably accessible location. IBM seems to be moving inexorably toward crazier and crazier keyboards, and the PC/RT "enhanced" keyboard looks to be designed for DisplayWrite and nothing else. I hope Gold Hill is listening and intends to clean up its act, because wedding its Lisp to IBM hardware seems like a disaster for anybody that intends to do Lisp development on MS-DOS systems. The more specific the system is to particular hardware, the less the incentive to develop the product in the first place. The MS-DOS market for software may originally have been able to aim at running on IBM PC's (and to h**l with the rest of the world), but that is no longer the case. Well-behaved applications are expected now and if a program is not well-behaved, it is not very likely to survive. If Lotus, Intel, Microsoft, Ashton-Tate and most of the other big hardware and software houses can (more or less) agree on things like expanded memory and the like, then Gold Hill should be able to make GCLisp at least reasonably well behaved. Yes, I know I have been flaming, but I think it is a d**n shame that GCLisp looks so untrustworthy. I have an application that I had been thinking of developing in Lisp, but I will probably choose a better behaved environment for it than GCLisp (possibly Prolog, maybe even TurboProlog), because I cannot afford not to be able to trust my development tools. -- {ihnp4,seismo}!{iuvax,pur-ee}!bsu-cs!nen j. I