student project

The problem is 1) recognizing the good guy, 2) admitting that he is worth keeping, and 3) that someone else is looking too.

I once had a recruiter ask (I assume for his client) what I would do if my current employer would match an offer from a new company. I replied that I wouldn't consider it because the fact of me talking to the recruiter meant that I'd already decided to leave (the current employer already blew the chance to keep me). I then paused a second and continued, however, that means any future employer can expect exactly the same.

It's hard to know who they are, too.

...and you can interface to any other language, calling REXX programs or vice-versa as needed.

I think that's about the best answer there is... after all, if you did stay, you're now on management's radar as "that guy" who was planning on jumping ship and hence probably will no longer be part of the "inner circle," advancement oppotunities will slow, etc.

You'll have to ask him.

--
For the last time:  I am not a mad scientist, I'm just a very ticked off
scientist!!!

Nice! Looks heavily Maxwelled, the stuff that scared the heck out of me back at the university. I am missing a bible passage in there though :-)

A client is rigging up a Beowolf cluster for sims. Initially I thought that was something from the old days but it's still done.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

I had to go through some code recently that has the occasional Spanish in there. Made me crave margaritas. Luckily we had burritos for dinner that night.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Thanks. The manual was written as much to remind me what I did (and what my plans were) as for other folks.

As I'm a consultant, perhaps the motto should be Leviticus 25:12: "For it is a jubilee; it shall be holy to you; you shall eat what it yields out of the field." ;)

Beowulf is one of those terms that gets fought over in the HPC world. You might suggest the Rocks cluster distribution,

I used Rocks 4.4 in my cluster at IBM, and really liked it. (The performance of my one lonely pizza box is better than my whole 2005 cluster, so I don't need Rocks so much at present.)

Cheers

Phil Hobbs

pants

Damn, you are really laying your prejudice on thick today. Really, accusing all programmers of being somewhat autistic is quite OTT.

You can't tell the difference between "most" and "all"? You must be a programmer.

John

John Larkin expounded in news: snipped-for-privacy@4ax.com:

He's just off by one.

Warren

What really shocks me is now many professional programmers, even embedded systems guys, don't understand basic math. We had a guy who had extensive realtime programming experience, but could never understand fractional notation, simple filtering, simple PI control loops, or manage to get ADC data into a 32-bit register without losing the scaling. It was easier to write the code ourselves than lead him through it step by step.

We spent the last few days cleaning out his office. It looked like his code.

John

Well! Larkin has become buddies with NymNoNuts, so I guess some of that asininity has rubbed off on Larkin ;-) ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
I love to cook with wine.     Sometimes I even put it in the food.

John Larkin expounded in news: snipped-for-privacy@4ax.com:

Sounds like he had bigger problems-- off by PI errors!!

Speaking of PI, I've always been fond of the programmer's technique for estimates: estimate what you really think it will take and multiply that by PI.

Warren

Aren't most programmers?

The productivity ratio among programmers is easily 20:1. But you don't have to pay the good ones 20x as much.

John

So speaks the man who boasted about an impossible convergence rate for his Newton-Raphson square root code starting from a fixed seed. Futzing with the numbers was the immortal phrase you used - the routine was declared as an example of "Larkin bug free" code at the time too.

Those in glass houses should not throw stones.

Sounds like you are not very good at recruiting programming talent. This could explain why you have such a jaundiced view about software engineering. There are good software engineers out there it is just that you seem to be unable to find them or pay the going rate.

Pay peanuts get monkeys is the golden rule. Sounds like you got a right one if they could not handle basic scaling. How did they ever get through the technical interview in the first place?

I usually find it is hardware engineers that have not thought through what happens if a project needs eg. a 14bit DAC instead of 12bits as a late change or later "improvement" for marketing specsmanship.

A tidy desk doesn't necessarily aid productivity. Some of the tidiest desks where I have worked belonged to the least productive people. YMMV

Regards, Martin Brown

An empty desk is a sign of a disturbed mind. ;-)

Cheers! Rich

It worked to the required accuracy over the possible range of inputs. I tested it to make sure. Of the thousands of boxes that used it, many tested by labs and customers, it seems to work fine, bug-free in fact. This was a 16-channel AC power meter, and I had to do 32 square roots, for RMS volts and amps, often. The upper limit of input was bounded by ADC saturation, and I clipped the low end to zero below some limit. That's a common trick in TRMS DVMs, so that a shorted input displays zero.

It's tough. You do the usual stuff: advertise, interview, look at code, check references. One problem is that a lot of programming is essentially qualitative, but our stuff is quantitative. Also, a lot of programming gets done by hack-and-debug, and we can't afford that.

The thing is, you really don't know much about a person until you actually work with them.

This

I know four superb programmers, modestly excepting myself, and employ one of them already. The one I work with is too valuable, as an architecture/electronic/FPGA designer, to have him spend all his time programming. The others have their own businesses far away and aren't available as employees.

We have good pay, big cubes, bonuses, 401K, medical/dental, indoor parking, nice kitchen, machine shop priviliges, loose work rules, a cabin in the mountains, and we are willing to go bail for spouses.

Sounds like you got a right

He seemed OK at the time. Turns out his personal life was crashing down all around him, which may have influenced his concentration.

We're pretty good at that sort of math and planning ahead for versions/upgrades.

The correlation is weak there. Electronics is messy, as is realtime programming of mixed-signal stuff: DUT, scopes, JTAG pods, logic analyzers, three or four monitors, power strips, lamps, ethernet switches, all sorts of cables and test equipment. I found one tangle in his office of 8 scope probes and maybe 20 ground clips. Whenever he needed one, he'd just go get another from the lab. No wonder nobody else could find ground clips.

Our current philosophy is to hire programmers as interns. That way, we get young people who won't fight us about the programming styles we want, and if it looks like they're not going to work out, it's over without hassles. If they do look like keepers, we can turn it into a regular position.

John

But a couple of square feet of visible surface is good.

John