Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!caen!spool.mu.edu!cs.umn.edu!kksys!wd0gol!rathe!orbus!gn From: gn@orbus Newsgroups: comp.sys.next Subject: Re: NeXT as a vanilla UNIX box (...and major disappointment (long)) Message-ID: <1991Jun23.001024.171@orbus.uucp> Date: 23 Jun 91 00:10:24 GMT Article-I.D.: orbus.1991Jun23.001024.171 References: <998@rosie.NeXT.COM> Sender: gn@orbus.uucp Organization: Society For the Prevention of Necrobovinism Lines: 73 In article <998@rosie.NeXT.COM> cmac@next.com (Chris MacAskill) writes: > I found this thread to be fascinating. It rekindled The Great Debate, at > least at my end of the hall here at NeXT. > ... > > In any case, our identity certainly isn't clear out there. I sometimes think > we've combined the planet's best products with less-than-the-best marketing I guess I just gotta get some stuff out of my system. We (Rathe, inc.) paid for a NeXT in January to develop a major railroad shipment tracking application. I personally spent alot of time convincing our client that NeXT was the platform to use. After seven weeks I had to go back to the client and reconvince them that the best platform for development was Sun. I had to do this because we had project time constraints, and contrary to several promises from NeXT we did not receive our order until the end of April. Necessary technical documentation until the first week in June. I knew things were bad at NeXT when three days after receiving the machine, I get a call from NeXT order processing informing me the system should ship at the beginning of the next week. The NeXT Rathe bought is now relegated to my home machine, and I've come to the conclusion this is for the best (for Rathe not NeXT, our client is the largest RR in the world, and NeXT could of sold alot machines into this market.) I believe this for several reasons: Documentation: The Documentation for the NeXT is at best sparse, incomplete, and in several case just plain wrong. They do not include printed Unix documentation. The calls section of the on-line doc is missing at least twenty system calls/ subroutines, there are commands included in the distribution that are not documented anywhere. There are nice features of Mach and NeXTstep that are also completely undocumented (shared libraries, class usage). Documentation, to me, is NeXT's biggest short coming. Unix: UUCP is broken. Signals are not quite right. Shipped Sendmail is brain dead. Security is weak. The worst, I feel is, dump/restore doesn't work, and neither does tar when dealing with multi-volumes (dump/restore has additional problems) The Mach implementation is also weak, missing several key features, like shared memory, and multi-processing. Hardware: Since receiving my NeXT I have had problems with: The Printer (flaky). The Disk Drive (replaced). and next week Motorola is coming out to replace the processor board (Panicking at least a couple times a week). Even the 2.8 meg floppy disk shipped with the NeXT has gone bad. The response I've most often received when speaking to others of my problems is that, indeed NeXT has problems, but who doesn't (horrors stories about HP & Sun follow) Doesn't Next sell itself, though, as better, and outside the circle of unix workstations? Better then those other guys. There are lots of things that are wonderful about the NeXT (price/performance, the user interface, Motorola repair, one of the best 'out of the box' bases of software and hardware features). This only waters down what are critical, major oversights and failures on the part of NeXT. -- Greg Noel, Rathe, inc. umn-cs!rathe!orbus!gn "Once performance is assumed, style is everything" -A. Noctor