Once or twice up here I've noted that I'm researching a history of the HP 3000 which I will write. This weekend I spent time catching up with a few of the better sources for 3000 histories — all while I was muddling through audio notes of my visit to the HP 3000 Software SIG meeting, held this summer at the Computer History Museum.
I'll have more to report on that CHM SIG meet, the first of which I hope will be many more. But for today, I'd like to point out a few bits of history the devoted 3000 user and devout HP Way worshiper might not know. For example, in Bill & Dave: How Hewlett and Packard Built the World's Greatest Company, Michael S. Malone includes a handful of pages about the HP 3000 saga. Malone has written a great HP book with just a dash of 3000 specifics. He notes that your system was the first product which ever embarrassed Hewlett and Packard in public — because it flopped so badly. The first 3000
...only supported two users, not the 64 it had promised. The computer sent was a pile of junk. It wasn't even finished, missing some key components. [Hewlett and Packard] heard about the dead-on-arrival 3000 from an article in Computerworld. It was the first piece of truly bad press Hewlett-Packard had experienced in 33 years of business, and the two founders... reacted ferociously.
It's good to recall history while trying to make some, like the efforts of the OpenMPE group to get a license to fix MPE/iX once HP has given up that business. Bill and Dave took measures in 1972 with the 3000 to ensure HP's reputation would survive intact. Some in HP will stand firm today to ensure no history can be written of a 3000 living and working beyond HP's involvement.
Every HP 3000 (called the System 3000) that HP had sold was recalled all during 1975, to be replaced by a Model 2100 computer for free if the customer wanted one. HP pulled the System 3000 off the market for the better part of two years.
It was bad enough to trigger an upheaval of talent. Tandem Computer — now the homeland of the HP NonStop server line — was born because of the HP 3000 flameout. James Treybig, Tandem's CEO, and his co-founder Mike Green came right out of HP's Computer SYstems division (that's the CSY you'll hear the 3000 group being called.) The two men founded a company that idolized the HP Way and opened up shop just down the street in Cupertino.
At the CHM meeting earlier this summer, a table of 3000 pioneers (the museum's title for everybody) talked about the scant uptime for the early 3000s. The computer which made HP a computer company, instead of just one of the world's greatest instrument and calculator creators, couldn't remain up for more than 24 days. Malone says the flaw was in the design, which overflowed a clock register. At the core of the problem was a gaggle of engineers telling HP that everything would be just fine with the 3000, instead of reporting what was well off schedule of being fixed.
Something very similar happened at the 3000's next debacle, when the server tried to become a Reduced Instruction Set Computer (RISC) and head toward 32-bit computing. HP responded in both instances with an enforcer. In the 1970s it was Paul Ely, feared by many according to Malone's book. In 1986, when PA-RISC was a 3000 bust, the muscle was HP's labs chief Joel Birnbaum, who said to us reporters in a conference at an HP user group meeting, "we expect that these problems will yield to engineering discipline."
Bad press prompted a fierce response once again, when the industry's journalists called out the woes of a long-awaited HP product. Like any good vendor, HP needs honest critique, told in public, to make the best products. If there had been only an echo-chamber with no stories, the 3000 probably never would have gone on to sell 200,000 units through the remaining 17 years of its HP lifespan, or become so durable that "between 10,000 and 20,000 of them" would still be running today, according to Allegro's Steve Cooper at the CHM meeting.
That's a message for HP to consider as it wants to control what can be heard and reported at the biggest meeting of its users. In a bit of irony, this summer Computerworld was being kept out of a Tech Forum roadmap meeting, along with the NewsWire. (We were the only reporters at the door.) The access problem was fixed, but I'm a little less certain about whether HP's intentions have been reset.
But it's useful to note that remembrance was not essential to this company's founders. Malone's book takes special note of the HP Way's devotion to sentiment, as in: None. Again from the book:
Hewlett and Packard had little nostalgia about the past unless it in some way enhanced the future. There were no sentimental stories about the HP 200A, the company's founding product [in 1939], by the 1950s. Once demand fell off, and all profits were wrung out of the product line, Bill and Dave jettisoned it with barely a glance back.
A generation later, the company would do the same with its most famous product, the HP-35 calculator. Had corporate publicist J. Peter Nelson not written, on his own initiative, a press release elegy of the device, its passing would gone unremarked.
Which brings us to a generation or two further along, and what seems like another kick to the curb for a foundational HP product. HP restored the founding garage in 2005 since the building was deemed the birthplace of the Silicon Valley by the State of California. Rebuilding a house and garage is one thing. Revering a product line is a step further, one HP's seems unready to take, or even chronicle.
HP Archivist Anna Mancini was invited to the CHM's 3000 meeting, but could not be counted among us. Perhaps at another gathering at the museum. But the HP Way was always dedicated more to profit as a chief objective, rather than undying devotion, as Malone says and the 3000's HP history echoes. A computer with enough history at HP to lead the company into the mainstream computer business — and from there to printers and PCs, and so to the Number One ranking among all makers — the 3000 didn't get treated any better than the 200A or the HP-35, once management calculated the 3000's growth days were at an end. And that too is the HP Way.