woolstar@cit-vax.Caltech.Edu (John D Woolverton) (01/27/89)
> Wonderful! Then maybe the features of 4.4 BSD can be merged into System V > release 9 sometime in the next century! :-) Just when their getting ready to merge Berkeley 4.3 into System V release 4.3 :-) >>What *is* sad is how much the third-tier companies are being constrained by >>what's happening on the first tier. Customers are demanding System V. Not SVID >>or POSIX or X/OPEN, but System V. If it doesn't look just like a 3B20, it gets >>shipped back. Innovation in UNIX is rapidly becoming a thing of the past. I was very unhappy with what our 3B20 came with: A c compiler that had numerous bugs. (I was surprised that any of the programs/system stuff worked) A bourne shell, cat, grep, find, and a few other /etc stuff. What it didn't come with includes: more, less, csh, ksh, telnet, rlogin, ftp, (networking ...), tip, emacs, man, uptime, who, finger, date, and all those little things that I like to use from day to day. One of the few blessing it had was that it had the software for the dmd 5620 on it so that I could actually have windows to it. (One of three people on campus who could do this.) For quite a while, it did get some use by a band of undergrads hacking empire on it. It was jokingly refered to cit-empire instead of cit-3b. One thing I'll give it though, when we had to remake all the empire code, boy would it race. As it is, we still do not have any form of network communications or hardware for it on our 3b20. woolstar@csvax.caltech.edu woolstar@cit-3b.caltech.edu :-) -- -------------- John D Woolverton "Yes it's true..." jdw@tybalt.caltech.edu woolstar@csvax.caltech.edu
vixie@decwrl.dec.com (Paul A Vixie) (01/29/89)
Anyone ported 4.3bsd to the 3B yet? The hardware ain't too bad... -- Paul Vixie Work: vixie@decwrl.dec.com decwrl!vixie +1 415 853 6600 Play: paul@vixie.sf.ca.us vixie!paul +1 415 864 7013
ron@ron.rutgers.edu (Ron Natalie) (02/01/89)
3B20's are not worth the bother. AT&T isn't supporting UNIX for them. The hardware, while suitable for a phone switch, is pretty lousy. The Ethernet interface doesn't work at all despite the $$$ I spent on it. The CE's are constantly grousing about having to lug these Delco batteries around and despite that the machine goes wierd everytime the power gets shut off anyway. The later processors in the 3B line actually claim to be UNIX machines and fare a little better. -Ron
martillo@cpoint.UUCP (Joacim Martillo) (02/01/89)
Anybody have any idea how much a 3b20/600 would cost?
vixie@decwrl.dec.com (Paul Vixie) (02/01/89)
# 3B20's are not worth the bother. [...] The later processors in the 3B line # actually claim to be UNIX machines and fare a little better. I've got to try to dig up Ed Gould while at Usenix. Depending on how hard he thinks it would be, I might be willing to *contribute* some effort toward a 3Bxx 4.3bsd port, just to see what the market for it would be. Comparing the documentation, training, and support you can get from MtXinu to what you can get from AT&T, it seems like MtXinu ought to own any market they choose to compete in. They've certainly stomped all commercial alternatives except Ultrix in the VAX UNIX market -- I have this wonderful fantasy of someone selling a supported BSD for the 3B boxes and putting AT&T out of the software business while somehow making it worthwhile for them to keep shipping iron. I know it's bizarre and could probably never come to pass for any number of reasons, but it's a pleasant fantasy. I once had to _use_ a 3B20: ugh! Disclaimer: opinions stated are my own, DEC doesn't know I'm posting this. -- Paul Vixie Work: vixie@decwrl.dec.com decwrl!vixie +1 415 853 6600 Play: paul@vixie.sf.ca.us vixie!paul +1 415 864 7013
dlm@cuuxb.ATT.COM (Dennis L. Mumaugh) (02/02/89)
In article <Jan.31.16.11.35.1989.385@ron.rutgers.edu> ron@ron.rutgers.edu (Ron Natalie) writes: 3B20's are not worth the bother. AT&T isn't supporting UNIX for them. Odd, I get paid lots of money to support the software (13 of us do) and others get paid to support the hardware. You did buy a support contract, didn't you? Our management would be interested in hearing about cases where your field engineers couldn't fix the problem. The hardware, while suitable for a phone switch, is pretty lousy. Good for what it was designed to do. It's really fast on I/O. Kinda slow for cpu bound stuff. 80 users isn't bad. The Ethernet interface doesn't work at all despite the $$$ I spent on it. Must be your hardware. I use TCP/IP all the time. A 3B20A is our major netnews and uucp gateway. We share the news via rlogin or RFS. The CE's are constantly grousing about having to lug these Delco batteries around and despite that the machine goes wierd everytime the power gets shut off anyway. Power off? Ours gets shut off only when Commonwealth Edison has a thunder storm and drops power totally. The later processors in the 3B line actually claim to be UNIX machines and fare a little better. Much, much better. The 3B2/700 screams but it still isn't quite as good at I/O especially terminal I/O. But the cpu is definitely much faster about 5xVAX or so. -Ron You could go back to good old PDP-11/70's Ron. The RP06's were real toasty during the cold winters. Seriously, you need to talk with the support managers as the 3B20 isn't such a bad box. Put 16 Meg of memory and System V 3.1 on it and life is bearable. Personally, I use it and my own private 3B2/400. -- =Dennis L. Mumaugh Lisle, IL ...!{att,lll-crg}!cuuxb!dlm OR cuuxb!dlm@arpa.att.com
ron@ron.rutgers.edu (Ron Natalie) (02/04/89)
Having a 3B20 inside AT&T is a lot different than being stuck with one in the field. When I was a consultant, I found that the 3B20 tty driver wasn't SVID complient due to a hideous bug in the VMIN/VTIME handling that caused processes to lock up in an unkillable state. I was working around this problem on an AT&T machine when they put up the new release of UNIX. A year and a half later, on the outside, I was still not able to get this release. As for the Ethernet interface. It refuses to work with anything but these black brick 3COM transcievers that AT&T provides. The CE's are worthless trying to fix this problem. Even when it is working the thing frequently goes on to net spewing continuous garbage datagrams. The telnet/rlogin response is pretty poor in any case compared to what we get on the 3B2's and 3B15's. I talked to some AT&T guys from a group in Princeton, and they admit that their machine is much more advanced with regard to engineering changes than ours. Unfortuatenly, even the $$$ I paid for the hardware upgrade to run TCP hasn't made a big improvement. My conclusion, AT&T really isn't exerting any effort into making the 3B20 a computing product. The 3B15's and the workstations have real support available for them. The 3B20 seems to be a lost soul. -Ron