lray@CIVILGATE.CE.UIUC.EDU (Leland Ray) (12/31/89)
>rtp1%tank.uucp@handies.ucar.edu (raymond thomas pierrehumbert) writes: >Jeez, so SR10.2 breaks NFS! There goes my plans for upgrading to 10.2. >I think I'll just wait and sit this one out for a while. Sure, I could >order the new NFS, and maybe it would work, but I'd have to cut a >purchase order, get it sent out, install the new tape, and hope it >works as well as what I've got now. Couldn't you have built some >backward compatibility into SR10.2? What else did you break? Will Welcome to the world of workstations, a world where the software grows and changes daily. Apollo likes to obsolete software; during the time I've worked with their machines they have changed binary formats twice, required that all disks be reformatted, and changed their install procedure quite a number of times. The amount of time I've had to spend patching and reconfiguring the system has grown. Of course, I don't really blame Apollo for that. While their software has molted, so has their hardware, so has their customer service, so has their sales, and so has their profits. They were well on their way to recovering when HP bought them. What I would do, if I had questions about the latest version of NFS (or any product, for that matter) is to find someone who has it, and get a copy of the release notes. After reading them, you can decide for yourself if the upgrade will help you. It is worthwhile to upgrade often enough so the people at the 800 number can help you. (Quiz: How many 800 number people know why an Imagen printer at SR10.1 truncates lines after 10-15 characters?) Just spendin' my days, Leland Ray Systems Administrator Soakin' in them cathode rays. UIUC - Dept. Civil Engineering (217) 333-3821
pha@CAEN.ENGIN.UMICH.EDU (Paul H. Anderson) (01/02/90)
From: lray@civilgate.ce.uiuc.edu Subject: SR10.2 and life Welcome to the world of workstations, a world where the software grows and changes daily. Apollo likes to obsolete software; during the time I've worked with their machines they have changed binary formats twice, required that all disks be reformatted, and changed their install procedure quite a number of times. The amount of time I've had to spend patching and reconfiguring the system has grown. In fairness, just look at Sun, who introduced a 386i, then announced plans to drop OS support for it and the entire Motorola line of machines (Sun2 and Sun3). This is an interesting alternative, isn't it? Besides, considering that I can still run binaries from three or four major releases back, things on the apollo aren't bad at all. I'm very impressed with how Apollo has always made strong attempts to preserve backwards compatibility, sometimes at costs that I would consider excessive. Source level compat has always been good, too. Look at the extensive effort to keep the DM alive in the face of X windows standards - at 10.2, both reside on the machine, either can be used independently, or a nice mix of both can be used with few problems. Most vendors have chosen to keep their manager and X completely separate, but not Apollo. Of course, I don't really blame Apollo for that. While their software has molted, so has their hardware, so has their customer service, so has their sales, and so has their profits. They were well on their way to recovering when HP bought them. I don't see that they were on the road to recovery, survival, perhaps, but not recovery to the top of the heap where they belonged. HP might change some things for the better. This note isn't intended to be an apologist's letter, especially since I'm having fairly extroidinary difficulty keeping our 500 nodes working. The bottom line seems to be that at least among the traditional manufacturer's, nothing really seems that great these days. Maybe DEC or IBM will surprise us with exceptional support for their machines, but I'm skeptical. Just spendin' my days, Leland Ray Systems Administrator Soakin' in them cathode rays. UIUC - Dept. Civil Engineering (217) 333-3821 Paul Anderson Apollo Systems Programmer Computer Aided Engineering Network University of Michigan
rtp1@tank.uchicago.edu (raymond thomas pierrehumbert) (01/03/90)
I never said the competition was any better, did I? I bought an Apollo, didn't I? Did I even say I wish I had bought a Sun or SG (I don't, btw)? Basically, Sun's whole product line has become as boring as IBM mainframes, except for the Sparcstation, which is a great little machine at a very fair price. All that notwithstanding, things in the workstation world are much worse than they ought or need to be, and if we are going to start improving things, why not start with Apollo/HP. I think that breaking something as important as NFS, and not coordinating the NFS release with the 10.2 upgrade, is very shoddy workmanship. If it hadn't been for the chance encounter here, I probably would have installed 10.2, found that NFS wouldn't work, and been out of business until it all got straightened out. Things shouldn't be that way. And I wouldn't mind so much paying the piddling $100 for a new NFS, if it weren't so hard to find out just what to order, and if it didn't take so long to get it. There is no good reason it should take longer to get a tape cartridge than it takes to get Mac or PC software, but it always does (see my posting on Fortran, below). If Apollo really wants to be at the top of the heap, they will fix this problem. And what ever became of the idea of producing a quality product and standing behind it? It is true that software prices (initial) at Apollo are about the lowest in the business, but the discounted present value of the maintainence charge makes the real cost astronomical. I'd rather pay a fair price up front, and then have Apollo stand behind there product, and guarantee that they will inform me when they do something to break it, (and to send along the appropriate update in a timely fashion).