Who is your favourite electronics guru?

That URL sends my browser into a spindizzy.

I've just been reading some papers by Tim Green of Burr-Brown. A chap worth following by the looks of it.

"Fred Bloggs" schreef in bericht news: snipped-for-privacy@nospam.com...

You can add yourself to the list, also not open for debate.

Hmm, yes. Interesting solutions to some problems. I have to sometimes sit down and work out exactly what he is doing.

Science has to do with the experimental refutation of hypotheses. There is no 'truth' - only a search for better theories which are themselves data compression algorithms.

Dirk

In message , dated Wed, 9 Aug 2006, Dirk Bruere at NeoPax writes

Well, perhaps not quite: 'mappings' might be closer. The data isn't deliberately compressed, but not all of it is acquired, and not all of what is acquired is used.

I suppose that technically a theory is a string that is the optimum (to date) compression of all known data. For example, Newtons theory of gravity reduces bricks falling, cups falling, apples falling etc etc, planets moving, cannonballs arcing, to a single simple string embodying the equations of motion.

Additionally, from Church-Turing there is no way we can prove that we have the string of minimum length.

'Truth' is simply shorthand for 'utility'.

Dirk

You mean how he thinks it works. Generally there could be several interpretations of the subcircuits and it is best if the original designer writes up his intent.

Most outsiders who know nothing about electronics assume that the practice is a mere application of technology, and actually quite a few jokes who consider themselves to be electronics engineers are the same way, thinking it is just a matter of locating the right IC or right app note. These people are of low quality and their work is garbage. I don't waste time dealing them, seeing major kinks, bugs, flaws in their so-called designs, I let them pass without comment or concern for the consequences. Then there are these other kinds of parasites who will take your design intended for one application and try to run with it in another, for which it is not suitable, usually at great expense, and then come back and blame you for their problems. This happened to me once, then an outside firm came in and proposed an improved re-design to reduce the cost from $15K to "possibly" $10K per unit. This was fine and dandy with me until they specifically called me out by name in the proposal and stipulated that I not be allowed to have *any* involvement whatsoever in the new development. It took me a few days, and then I came back with a counter proposal for a radical design change that reduced the cost per unit to $0.5K. The outside firm was soon notified that their services were no longer needed, and their manufacturing contract for production of indefinite quantity of the $15K/unit items was terminated.

My only help comes from a fairly massive collection of textual material, some of it going way back. I even have a musky copy of that Analysis and Design of Integrated Circuits to which you contributed chapter 15 IIRC. I know all about you, your work, and your capabilities. That SED collection on your site is at the low end:-)

Should have sent you to

Sounds a little too simplistic to me... for one I don't think it's a valid conclusion that Ro is resistive.

I'll try it with a real OpAmp (device level) and evaluate.

...Jim Thompson

Oh, Fred! I'm so agog! You're just so-o-o-o wonderful ;-)

...Jim Thompson

You've both touched on extremely important points of epistemology...

The first is basically Sir Karl Popper's point that a hypothesis is something we set up for possible refutation, and a successful hypothesis is merely one that could have been refuted but wasn't. I think Popper would say that there is such a thing as truth, but we do not apprehend it directly (in non-trivial matters).

The second one is very important to anyone who does artificial intelligence. Any intelligent agent, moving about in the world, doesn't want to acquire

BTW, the word "true" in my second paragraph should be understood as restricted by the first paragraph, i.e., "probably true" or the like.

I think there's a lot more to it than that. As long ago as Aristotle, a distinction has been made between laws of nature and generalizations that just happen to be true (no matter how true they are), e.g., there are more cars traveling north than south on this highway.

Yep. When I was 25 years old.

Perhaps. Two reasons...

Much of my work is covered by NDA's

The rest is art which I will never divulge until my death-bed book is released ;-)

...Jim Thompson

Don't be embarassed. We all did silly things when we were kids. I did all that C-5A and S1B moon rocket stuff for about $6 an hour.

John

In message , dated Wed, 9 Aug

This is indeed very important. There may be 'conjecture' before a hypothesis emerges, and only when refutation fails (for a sufficient time) is there a theory. But try telling media people that a theory is something tested and reliable, not 'only a theory'.

We don't know how to make positronic brains (© Isaac Asimov and Susan Calvin) yet. We do know that what we think we see at any instant is a great deal more than is actually the case; the brain concentrates on a much narrower field of immediate attention.

Of course, traditional robots have far better memories than humans, and this is probably very easy to achieve.

In message , dated Wed, 9 Aug 2006, Jim Thompson writes

The second paper deals with that.

I'm not embarrassed. I think it was rather good for a kid barely out of school.

I was making $8.65 an hour ;-)

...Jim Thompson

In message , dated Wed, 9 Aug 2006, Fred Bloggs writes

And you such a charming fellow, who suffers fools gladly, too! How unfair. (;-)

They were motivated by FEAR - fear that you would do what in fact you did. I've just come into a project where something similar has happened. The main client has, after wasting 3 years having a custom system developed, finally given up on that 'expert', and is now persuaded to use a commercial system that costs peanuts and probably does the job quite well. I have to find out whether it does or not.

Hey, isn't it about time to rediscover the square-wave divide-by-three circuit?

John