August 13, 2011 at 8:57 am #2210911
No, Not Confused, using examples outside of PC businessLocked
by lsb4000 · about 11 years, 1 month ago
article rootThis conversation is currently closed to new comments.
August 13, 2011 at 8:57 am #2843810
by lsb4000 · about 11 years, 1 month ago
You certainly have an interesting view of what happened when IBM formed GPD (not GTD). My view is very different. My dad was involved in the planning tha led to its formation. In my view it was formed because IBM felt it needed a more nimble organization to address the needs of small businesses. It had little-to-nothing to do with the government lawsuit even though there was internal talk about an eventual split of the company.
I joined GSD when it was formed and later went to GPD HQ, only to have my project sent back to GSD and Boca.
Unfortunately GPD (mainly GSD) did it to itself by allowing multiple competing architectures make it to market. At one point we had 8 different architectures in the market, confusing everyone. Yes, many were good architectures, but not very good business. The AS/400 was an attempt to consolidate the System/3, 23, 32, 34, 38 and even the Series/1 under one umbrella. The Series/1 part didn’t make it (as you alluded) but the AS/400 has been very successful,
Unfortunately, the multiple architecture thing really hurt us in the minicomputer/engineering/process control market. I don’t know where System/7 came from. I think it was a bunch of DOS/360 guys thinking a machine connected to a host machine was the way to go. The hardware was great but the software was from hunger. The Series/1 was much better planned, but it was a little late to the market and the momentum to take over that industry was largely lost. Estridge’s RPS was a real bummer so the sales division came up with EDX, which had been very successful on the 1800. Most of the Series/1 sales were based on EDX because RPS was the wrong OS for a minicomputer. The Series/1 did okay in spite of this, thanks largely to Special Bids.
The 8100 was indeed built on the UC. Having been on the UC design team, I know it was designed to better suited for a Point of Sale controller. I gave it the test and set instruction that allowed for multiprocessing but never envisioned anything like the 8100. The UC was as underpowered for the 8100 as the 8085 was for the System/23 (5120 by the way).
Thank you for reminding me of the name of the 5100/5110 PALM processor.
I could go on and on about this because I had to live it. I finally left after having gone back to DPD, being frustrated in trying to get DPD to sell GSD products.
August 15, 2011 at 9:41 am #2843706
Certainly we are contemporaries
by sensor guy · about 11 years, 1 month ago
The “nimbleness” cover story is very true. The folks that built the then planned future corporation were told that. My source was the executive responsible for the legal archives during the anti-trust suit and one of the technical people the lawyers had during the pre-trial hearings. The technical resource went on to become a VP in SCD, then fell from grace and became my first line a decade later.
The AS/400 was a processor fix to save the investment of Fort Knox. When they opened the doors and found snakes instead of gold, then they went to bifurcated CECs with identical I/O to save the money invested. Thus Olympic and Voyager were born. Olympic was destined to merge (successfully I might say) the S/3, S/3x lines (the S/23 never sold much except in Costa Rica and Cuba, but that’s another story for another day) and the Voyager went to merge DOS/VSE, VM and MVS as well as an attempt on DPPX called DPPX/9370 (which was announced). DPCX and MUMPS were killed.
The Series was never really seriously contemplated after Fort Knox. RPS (which I knew well because a college classmate of mine was its lead architect) was essentially a copy of the 360/44 manuals under VSE copied to run under a Series/1 processor, which failed miserably. EDX did not come from 1800, it came from System/7 and it came from sales. Walt Grain and Ed Bruklis from Boca were instrumental on that deal with a sales guy named Moody Steadham. I helped with the EDL command syntax. EDX sold well at Sears and State Farm (SE was Linda Lisak) and RPS sold well at Seibel’s-Bruce because of their SE, Ed Heaton.
The System/7 was SPD’s answer to the Raleigh SCD monopoly of the 3705 which had the anti-trust legal leaders blessing . They wanted a programmable SNA PU4 unit and the US Navy wanted a sensor controller for their ASW systems and that made the S/7 be born. It also allowed the DEC install base to be added to the anti-trust case. The PU4 plan went down in flames, but because of the lawyers they kept on with the sensors and later as a programmable communications PU2 (3274) or BSC (3271/3274) controller for ASCII devices, which the S/1 follow-on did well when we had our first supply chain fiasco with the 3270 display in the late 70’s.
I was the guy who convinced a major IBM customer to sue IBM (very long story) and part of the settlement was the elimination of loop on the 4680, moving them to Token-Ring and later to ethernet. RSD was stuck on loop after they got it to work, and SPD was only happy to make money on the detritus of Orbit-Neptune/3790-8100.
The 5120 and S/23 are different machines. The 5120 was the last PALM processor box. Here’s the two links for them:
Rochester borrowed a processor number assigned to the S/32 team for the S/23 and used it for the short period the box existed, although it sold very well under the new GSC under Akers-Gerstner in Latin America (especially Cuba) when we started to figure out how to bypass the US and European tax, accounting and import laws, but that’s another VERY long story.