Write-only memory     Real Innovation . . .

How many of you remember ole Bob Pease

No AI in those days :-)

Oh, yeah, Pease, and Widlar and Williams, the wonder trinity back in the day. I especially liked the way Bob responded to comments, submitted by snail mail, of course; he would scrawl his thoughts in the margins, between lines, wherever whitespace lived, and return the whole thing - no fuss, no muss, nothing to tend or throw out later.

Pretty smart group, and Bob Pease was easily the most approachable. Had some good arguments with him about his view of the limitations of 8 bit microprocessors, and especially when he got going about fuzzy logic and neural net tech.

Since we were already fielding products that used both, and 8 bit machines, it wasn’t hard to make an intelligent protest. What we never got around to resolving was the larger view, in which he undeniably was correct about the (theoretically, quantum limited) superior resolution of analog solutions.

Since I just spent an enjoyable evening listening to vinyl recordings on the new turntable my wife bought me for my birthday, it should be clear I wholeheartedly agree with his view on this and many related matters, like the likely lack of unique utility of the plethora of ‘high end’ audio cables of whatever kind.

He agreed that the one area of quality in which ‘high’ might be useful is the mechanical engineering side of the equation: if the thing won’t hold together, it’s no good.

But the electrical properties claims don’t stand up to scrutiny, and he mentioned this a lot. As for the ‘high fidelity’ claims among vinyl aficionados, well, they do have merit, again, purely from the standpoint of physical structure. But to gain access to that (somewhat subjective) experience, it’s only necessary to understand what one needs to translate molecular level resolution into a reasonably accurate reproduction of the information contained:

A carefully produced recording, not the simplest requirement, nor the easiest to predict;

A reproducing transducer capable of harvesting a reasonable amount of the information contained, without overwhelming distortion or noise

An electronic processing chain that doesn’t add its own information obscuring distortion and noise;

And finally, a reproducing transducer that can present the processed signals as sound without too much of their own interference.

We do all this with much less than anyone obsessed with fidelity would recognize: the new turntable (a reasonable design, with a reasonable cartridge, well thought out to do pretty much only what we need, and therefore, not too expensive, ~$500), an amplifier with relatively low power, but the redeeming characteristic of intentionally avoiding, by design, the gaffs that trip up the good intentions of the front end of the signal chain (it’s digital, something Bob would have laughed at, and mainly monolithic, to boot; it’s not your usual digital amp solution, a kind of sleeper product, patented, misunderstood, now the technique, and the company, mostly long gone)(it’s Tripath, check it out, doubt our claims of sincere adequacy if you wish); finally, small speakers we built ourselves, an inspired design (‘Overnight Sensations’) by a smart guy with no ax to grind, yielding a serendipitous conjoining of two pretty good drivers and a clear and honest understanding of audio reproduction and measurement thereof.

The amp and speakers cost half of what the new turntable cost. Good vinyl is pricy these days, too, and not a lot of what we might want to listen to is available out there.

But this rig is wonderful, and we have exactly the necessary ears to vet that. In fact, if Bob were here, it would be the answer to his oft repeated challenge, i.e., show me something I can hear, or go away. We’ve listened to two discs so far (new turntable, yes?) one a classical recording with which we are intimately familiar (Orff, Carmina Burana, Ormandy & Philly) in a venue we know well (Academy of Music), the other a jazz recording by our younger son, playing a distinctive Selmer Mark 6 tenor with a distinctive mouthpiece, a very high quality recording that I know started out purely digital just after the mics, but clearly, at a standard that vets Bob’s argument against most common formats, like the 16 bits / 44 KHz of a standard CD.

We can further vet his argument against silly claims of ‘golden ears’ that can hear things that scientific engineering analysis can’t explain; we do this with ~$750 worth of stuff, nothing extraordinary, just the right stuff at every stage of the reproduction, meaning as good as it can/needs to be, so, good enough.

We wouldn’t bother with fancier equipment, because there wouldn’t be a payback in terms of quality. We could change the requirement, and plug a whole different set of solutions into that, for a lot more money, and again produce something we could have demonstrated to Bob with certain success.

But such a ‘better’ system would still fail the practicality test; when we have reproduction accurate enough to identify a specific tenor sax with which we are intimately familiar, or spot an approximate location of the reverberation profile of a large, iconic room like the Academy of Music, we’ve jumped way over the limitations of the digital (CD) versions of the same materiel.

And we’ve done it in a way that fits exactly the way Bob looked at these sort of problems, although he likely would have correctly doubted some of the details until we were able to demonstrate their utility convincingly. 

If we couldn’t, we would have been the first to admit it, and start to look for what we thought we were bringing to the table. Another Pease / Widlar / Williams characteristic, in my experience; I think all three would approve of some of my own work, like the data system with a 2^26 dynamic range (that used AD parts, ‘cause the comparable BB parts lost the contest; sorry, Bob)(he was long with National when this happened, but still...)

He’s missed; they all are, I think, and if that makes me a dinosaur, well, they’re still using a few of those 300 or so 2^26 data systems out there. Bob Pease is a major part of the reason we could prevail at a design task like that, a kind of beacon of honesty and practicality, to a fault that I still find admirable.  —T

Thanks for your thoughts and reminiscences, Tom.

Thank you, for stirring up old memories...

I never saw that data sheet; it predates my earliest kidding around about WOM, which was triggered by a real architecture classification that didn’t last, Write Once Memory. 

Write Once Memory was just a fancy way to do what fused ROMs did, apparently a subsystem designed for hard storage of important data back in the days when results were a lot less certain to be available tomorrow, or even later today.

It was brought up in a meeting with some real old timers, part of the senior team when I was first starting out. 

If I recall correctly it was associated with some early minicomputer, maybe the SDS, or perhaps an even older system, none of which I ever actually laid eyes on.

All of these early systems had tons of ad hoc, impractical acronyms, most of which we promptly forgot. When we heard ‘WOM’ our brains immediately changed the ‘Once’ to ‘Only’, and a mini-meme was born.

We slipped a WOM in on an architecture presentation once; it raised a question, but I knew the guy asking had no idea what we were talking about, so I played it straight, followed the connections on the slide with an early incandescent powered ‘laser’ pointer while I completely overlooked the fact I wasn’t saying anything about the function it performed in the system.

I flipped to the next slide before I finished my last sentence, jumped right in, and had to stop as the guy broke in. He was from home office engineering, way above my pay grade, a whisper from being a VP in a $4.5 billion multinational corporation; I had to listen, but I was kicking myself for not fighting the urge to ignore his question, or be quick and vague, the usual secret handshake that signals I don’t know what it is, either.

That would have worked, but it was too late now. Imagine my surprise when he half stood, looked around the table, and took a moment to compliment me for my candor and willingness to communicate technical details to those who don’t necessarily grasp them in terms they can easily understand.

Yikes! I think they thought the red face was a modest blush, or maybe it wasn’t so noticeable with the lights down. I plunged on, finished my part of the presentation as quick as I could and sat down. 

When we were leaving one of the other home office engineers (a buddy who was as sharp as any of us in my Instrument Development Group) caught up to us, gave me a big grin, and half whispered “Nice recovery. Now he thinks you’re a genius.”

That was even scarier. I was never certain he hadn’t meant it to poke at me, ‘cause I know he knew I was totally BS’ing the guy. I worried for a few months that I’d run into him and be stuck acting like he understood the material we’d given him, but it never happened.

Still, I’d like to think that somewhere, that future VP was bragging to his buddies out on the golf course that HIS guys were using WOM in THEIR latest designs...    —T