A better approach to desing may be teachable.

a paper that is probably on-topic in a design group.

pdf

working out new design has been ripping up what initially looked like promi sing approaches and starting over.

hat they talk about as a strategic mind-set - more emphasis on where you wa nt to end up than on any particular way of getting there.

ut this subject, or any other subject for that matter.

ion toward strategic behavior during goal pursuit."

nts, really overdoing it with authors too stupid to learn the applicable ex tant science, so they just make stuff up as they go along.

emy of Science. It's a fairly high prestige general science journal - there 's nothing remotely phony about anything that gets through their peer-revie w process, and that does include checking that the authors have made proper reference to the existing literature.

lf this way.

tal health problem. Phil Allison thinks that everybody else is autistic, wh ich may reflects the same kind of problem.

he efficacy of masks in preventing infection. I guess it escaped you they w ere using data from influenza studies.

against virus-loaded airborne droplets, and the nature of the virus is irr elevant to their efficacy.

beat even that.

Not to what the mask does. If the virus is more infective, you'd be unwise to to rely on a mask alone, but nobody is recommending that. The mask is ju st one more item in the armory.

frequently bizarre.

--
Bill Sloman, Sydney

It is not that managers are bad (in my experience) it is that (especially program) managers frequently score success wrong. Many misguided managers can be steered into doing the right thing it just takes more effort

It's called 'tail wagging the dog'.

RL

I think that management, as taught by academics and institutions, attempts to codify and regulate and schedule design, a process that needs to be erratic and chaotic it critical stages.

Application of the sunk-cost fallacy, and groupthink, and procedures, prevent new ideas from displacing bad ones.

I work with one company that takes 2-5 years to design and qualify an electronic box, something we'd do in a few months. They have a documented WoW (Way of Working) that defines SIX versions, from breadboard to system deploymen, and they always do all six. We just do one.

They deploy, in systems, boxes that don't actually work at all. The WoW doesn't prevent that.

--

John Larkin         Highland Technology, Inc   trk 

The cork popped merrily, and Lord Peter rose to his feet.   
"Bunter", he said, "I give you a toast. The triumph of Instinct over Reason"

a paper that is probably on-topic in a design group.

df

orking out new design has been ripping up what initially looked like promis ing approaches and starting over.

at they talk about as a strategic mind-set - more emphasis on where you wan t to end up than on any particular way of getting there.

y program) managers frequently score success wrong. Many misguided manager s can be steered into doing the right thing it just takes more effort

There's always the delusion that the process is predictable. It's always an exploration of possibilities, which is frequently erratic - having to tear up up you've got so far and start over does slow it down.

"Chaotic" isn't a necessary part of the process. You usually know where you are going, though you can't be all that confident about whether the next s tep is going to take you further rather than up a blind alley.

Not understanding what's going on - or not being willing to sit and listen until you do understand it - is the basic problem. Substituting buzz-words from a management course for actual information is one way of doing that.

If he could do it all. John Larkin doesn't seem to go in for big and compli cated projects.

om

That's simple incompetence. Having the right procedures doesn't help much i f the engineers don't know how to execute the procedures to get the require d result, and corner-cutting managers have been known to try to save time b y skipping steps.

My electron beam tester project ran from the middle of 1988 to the end of 1

991, and during the first year the first project manager kept on "saving ti me" by skipping design reviews.

It wasn't until 25th January 1991 that we found the last of the bugs in sig nal processing.

"The problem is that when you down-shift a negative 2s-complement number, y ou have to back-fill the high bits with "l"s, and we back-fill with "0"s fo r both positive and negative numbers."

We'd sub-contracted the detailed design of that board, and back in 1989 the then project manager had taken the design straight from the subcontractor to the printed circuit layout sub-contractor without letting us (and me in particular) review the design. I can't be sure that I would have caught to drop-off then, but it was a pretty obvious conceptual error. As it was, I n ever got the chance.

The same circuit design had some 500 pull-down resistor on ECL outputs, and I'd had to spend a couple of days in 1989 going through the circuit diagra m getting them right - the surface mount assembly shop had noticed that the parts list didn't tie up with the circuit diagram, because we hadn't done a design review on the layout either - I'd seen the layout while it was in progress, and got the layout draftsman (who was very good) to polish up bit s of the transmission-line tracking, but we'd skipped the go-through-every- last-connection check that I did whenever I could.

--
Bill Sloman, Sydney

It reminded me that a lot of what I've gone through when I've been working out new design has been ripping up what initially looked like promising approaches and starting over.

It also depends on the size and historical success of the organization. At IBM's East Fishkill fab, before 1992, there was a group of engineers who seemed to specialize in throwing rocks at whatever new ideas were proposed. After the company's near-death experience that year, they were all gone.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

If you have the time, spend a week thinking and scribbling and filling trash cans. In a group mode, have a few whiteboard sessions. Use background brain power.

We know how to do this, but newbies can be taught. It's a shock to them that The Masters start out confused and don't proceed confidently along a decisive path.

I've had potential customers who really got into the initial exploration of possibilities, and other who thought we were incompetant amateurs for not knowing exactly what to do, in PowerPoint yet.

Intel has Copy Exact, which is the formalization of rock throwing.

There does have to be weeding out of truly bad ideas, but it shouldn't be done too soon.

Has anybody written a book about the mental and social processes of electronic design?

--

John Larkin         Highland Technology, Inc 

Science teaches us to doubt. 

  Claude Bernard

df

If you are doing anything demanding - and John Larkin seems to avoid that k ind of problem - you waste a lot of time if you don't spend the first week or so thinking and scribbling and filling trash cans.

There are always nice simple seductive approaches that turn out to be wrong , and it's a lot cheaper to find this out by thinking about them rather tha n by turning them into complete circuit diagrams, or breadboards, and findi ng out that they don't work - and never could hvae worked.

If there's any around.

Or think they do.

It takes a while, and some people are much too proud of what they think the y know to notice that it isn't enough.

d confidently along a decisive path.

It shouldn't be. That is the human condition.

People who would know how to spell incompetent ...

It all depends on the problem. Standard problems do have standard answers.

Perhaps.

Perhaps. It's a fairly small corner of the design universe, and a bit easie r than designing things like bridges, catalysts, drugs, vaccines and opera houses.

--
Bill Sloman, Sydney