koko: exploring NEXTSTEP 486

“Genghis Khan and his brother Don
Could not keep on keepin’ on”
(1971) “You Ain’t Goin’ Nowhere” – Bob Dylan

installing NEXTSTEP 486 3.2 on JAWSWorking with the JAWS machine brought me back to thinking about NEXTSTEP 486 being developed on that machine. I only know part of the story first hand. Andy Groves and Steve Jobs were friends. Between the two of them, it was decided that NeXT should create a 486 version of NEXTSTEP and that JAWS would be the best development platform. If I recall correctly, it was late December 1991 that Steve, Avie Tevanian and a few other NeXT folks came to Michael’s conference room and told Michael, Glenn Henry, probably Dennis Jolly, some other Dell folks and me about their plans. Glenn and I went to hear Steve announce NEXTSTEP 486 and demo on JAWS at NeXTWorld in January 1992. Dennis was the primary sales VP providing impetus for joint Dell/NeXT efforts — Dennis and I would routinely go to the (in)famous NeXT headquarters in Redwood City to discuss our joint plans.

NEXTSTEP 3.2 for the 486 was released about the time I left Dell in October 1993. Somehow I received a box or two with the installation media and documentation. For 25+ years it sat unused as I occasionally wondered if I should try NEXTSTEP 486 on my JAWS machine. Finally, I did.

It took quite a few tries to get NEXTSTEP 3.2 working properly on the DGX. While Dell SVR4 restricted support to a subset of the peripherals Dell sold at the time, NeXT tried to support a very broad selection of peripherals. While Dell SVR4 successfully attempted to autoconfigure for most peripherals, NEXTSTEP 486 depends on the person installing the peripherals to be knowledgeable about IRQs, I/O addresses and peripheral memory addresses. Once I was an expert at such things and am still able to sort things out adequately for NEXTSTEP.

NeXT Mandelbrot demo on JAWS

NeXT Mandelbrot demo on JAWS

NEXTSTEP 486 also seems to have a serious Y2K problem, or maybe just problems with dates in general. The installer considers dates much past 1993 to be “preposterous” and resets the date before proceeding. Once I set the date to the current date after install — the kernel panic’d on reboot and I was forced to reinstall again.

Enough people have griped about Netinfo that I will limit my griping. Though I still have a couple of PowerPC Macs that run Tiger (OS X v10.4), and, in principle, could provide Netinfo service for NEXTSTEP 486, I had no desire to go that route. NEXTSTEP seems to allow for NIS, as an alternative to Netinfo, but I had no desire to setup NIS, either. I finally figured out how to get the networking to work relatively normally without Netinfo or NIS, and get NEXTSTEP to boot uninterrupted with complaints about missing Netinfo/NIS servers, but the trial and error was no fun, and it is still challenging to make simple network changes.

In any case, I was thrilled to have NEXTSTEP 486 working.

P.S. NEXTSTEP seems to not have anticipated large drives very well. Besides the IDE/PATA problem I’ve discussed before, I haven’t gotten NEXTSTEP to install on disks larger than 1G. I can get NEXTSTEP and Dell SVR4 to dual boot on 1G drives, so I was hoping to have both available on the 9G drives I’ve been using with Dell SVR4.

2 Responses to “koko: exploring NEXTSTEP 486”

  1. Technologists Notes » Blog Archive » koko: reviving timbl’s WorldWideWeb browser Says:

    […] With working NEXTSTEP 3.2 in hand, I wondered if I could find sources for the original browser that Tim Berners Lee wrote when he invented the World Wide Web at CERN, and if I could, could I get that browser working. Success! […]

  2. Technologists Notes » Blog Archive » NeXT, give Steve a little credit for the Web Says:

    […] koko: exploring NEXTSTEP 486 […]