Sudden Confusion

hell is it

non sequitur.

The question is of giving other peoples' money to the poor of your choosing.

Who said I'm not helping people? I'm averaging a thousand hours a year these days, and 20% of my much-reduced income.

I'm just not robbing others to pay for it, like you want to.

-- Cheers, James Arthur

SMA

The literature is mostly about FFT/divide/reverseFFT algorithms, and then goes off a limb tuning them to avoid noise explosions in general and divide by zero dilemmas in particular. I read up on all that, and figured there must be a better way. There is.

You have such a dogmatic respect for "the literature." And you seem to do nothing much yourself. Anybody can copy algorithms from "the literature", and you apparently think such copying is somehow superior to understanding the process and inventing something new. How very weird.

That may be right if you factor in all the borrowing. But rates are so low, borrowing is basically free. Short term (90 day) rates look like

-Bill

That rate's artificial, an illusion produced by the gov't buying its own debt. It's a re-election-scam.

"The Federal Reserve is propping up the entire U.S. economy by buying

Long-term Treasury rates average ~5%, IIRC. Today, that'd be $800 billion a year on the gross debt.

-- Cheers, James Arthur

Forgive me, I do not understand.

Does this mean that the Federal Reserve is a gov't agency ??

I am sure I missed something, but what ??

hamilton

...

Obama has to borrow money constantly to pay the bills for all this spending.

The Treasury isn't able to find people (or countries) to loan us money at rates we can afford. So, they're colluding with the Federal Reserve. Treasury issues bonds for the money they need, at interest rates they like, and the Fed buys those with newly-created "dollars," to buy the bonds that no one else will.

Eventually we'll have to borrow at real market rates, and interest expense will leap accordingly. That will be ugly.

Obama's solution? The Buffett rule, $5B a year, plus spend more. Oh, and stop oil and coal. Underwhelming.

James Arthur

xes

tax

so

We have more needs than ever before, yet GWB left us broke between his tax cuts and two unnecessary wars. What to do?

Average income (constant dollars) for the top 1% has tripled since

an

, where SMA

ce

rts

ou

l,

the

ll

Scarcely. I'm the one with a history of publishing rude comments in Rev.Sci.Instruments pointing out that particular bits of the literature are inadequate. On the other hand, science is about improving our collective understanding of the world by publishing our various understandings of what's going on and building on the best of the earlier work.

My "dogmatic" respect involves reading the literature critically, and forming my own conclusions about what makes sense and what doesn't. Since you don't seem to have mastered the art of reading critically, this would seem to be a closed book to you.

At the moment. My 1996 paper is the only one that has generated a significant number of citations, and 14 isn't exactly a world-beating number (my wife's best paper is now up to some 800 citations), and my most recent publication - another critical comment, in this case following up the 1996 paper - came out in February 2011, though the comment was submitted (after a little discussion here) on the 21st September 2010.

The current project is beginning to look as if might lead to a publication in a some months. The little test transformer I wound last week and put in the post to London still hasn't got there, and the LTSpice simulation based on the John Chan model of an inductor with a hysteresis loop is still running - after ten hours (15msec of simulated existence)the model is still in a physically unrealistic state, and it's showing a depressing lack of enthusiasm for evolving into the physically realistic condition in which I'm interested. The simulation does run faster when the various components involved don't spend an appreciable part of their time in saturation, but at the present rate of progress it would take five days for me to get the

Copying a known algorithm obviously isn't superior to inventing a better algorithm, but it's easier and quicker, and the process of establishing that a newly invented algorithm really is better than the competition is also time-consuming and subject to error.

Publishing your invention is a way of exposing it to critical comment. It puts your competitive edge at risk, but can save you from investing in a lemon. Since you don't seem to be able to manage self-criticism, and don't publish, you are reliant on the critical skills of your employees, who won't be be all that enthusiastic about telling you that you've screwed up if your reactions here are anything to go by.

-- Bill Sloman, Nijmegen

e:

=A0It's all a

ism of

no need

GDP.

en

A point that they spell out directly below the piechart. They admit that their choice of 80% is controversial, and admit that other commentators opt for between 50% and 60%.

I can recall you pulling much the same trick - but attributing all of the interest on the national debt to the social security budget - and being rather less honest about what you were doing. Claiming that Johnson's Great Society spending represents a $16T expenditure that should all be added to the national debt is an equally arbitrary assignment.

It ignores the fact that Johnson's Great Society expenditures were an investment in the less well off, and paid off in a generation of more productive citizens, but since you don't believe in that kind of investment you are ill-equipped to estimate it's real costs and benefits.

-- Bill Sloman, Nijmegen

xes

tax

so

Actually Obama's solution is to get the US economy back to running close to full capacity; he can then taper off the stimulus spending and the tax take will rise as everybody starts making more money.

James Arthur's ideological blinkers blind him to that aspect of US Federal spending, and his point of view would see the US in 1933, at the bottom of the Great Depression, as in an ideal condition - GNP was down to 75% of what it had been in 1929, and unemployment was running at 25%, but the free market was perfectly free to optimise the economy (which it obviously wasn't actually doing, but James Arthur can't perceive that).

Various right-wit nitwit governments in Europe are making much the same mistake - driving their economies into recession to keep the budget deficit below 3%, which cuts the tax take even further and forces even more economy-damaging economies.

The current Dutch coalition government seems to have fallen yesterday when the silliest right-wing nitwit component of the coalition refused to sign on for the next round of expenditure. Geert Wilder's is a shameless opportunist, and seems to have decided that electoral damage caused by associating his party with the latest round of economies is going to be worst than the damage caused by his acting as a self- interested jerk.

-- Bill Sloman, Nijmegen

where SMA

No. The FFT based inverse convolutions that I found in the literature would have been a nightmare to code and tune, and if my results resembled theirs, not very satisfactory at that. HP, at least for a while, used the Bracewell Transform (for which I got a copy of Bracewell's book) to clean up/normalize TDRs, but it didn't look much better to me than the complex FFT divide. My algorithm is a few lines of Power Basic, involves no transforms, has no divides, works beautifully, and is fun to play with.

What's cool is to try to over-ehnance a rising edge. You can, say, take a clean slow rising edge and FIR-filter it to improve the risetime by almost 2:1. But add a little noise and push the risetime goal a bit, and it explodes spectacularly all over the screen. As do the FFT/Bracewell versions, for the same underlying reasons, but the mechanism is totally different. Most cool.

Why do you assume that I'm as lazy as you are?

where SMA

Well, it was worth it just to get a copy of Bracewell. Excellent book. I took Fourier from Bracewell himself, back in about 1984. One of my favourite classes, and hands down the most useful one I took in grad school. (Statistical optics was cool too.)

All the huffing and puffing that goes on over "ill posed problems" and the horrid dangers of deconvolution are way overblown. All you have to do is keep an eye on the noise gain, and not try to get something for nothing.

Cheers

Phil Hobbs

What will happen, certainly after the election, will be massive inflation. That's what governments do when they can't pay their debts: print money. That is equivalent to reneging on debt and stealing everyone's savings.

It would be especially cool if the US and Europe did it simultaneously.

We don't need hyper-inflation. Numbers like 2:1 or 3:1 will do.

You're right that we don't worry too much about parts cost. The high-performance parts are expensive, but we need them so don't compromise much there. Connectors and opamps and resistors and FPGAs are cheap enough that we don't spend a lot of time on them. Other things overwhelm parts cost, like the cost of specifying and purchasing something new, and the cost of kitting and placement and inspection. It could cost a dollar to add one 1-cent resistor to a board, if it has to be purchased, binned, kitted, and placed. So we lately use a lot of R and C networks, and try to keep the number of lines on the BOM down, and use what we have in stock, which all increase design time and design risk. If you use a resistor 4-pack, you lose the option of changing the values independently, so you have to be very sure it's going to be right.

I wrote a little program to prowl the resistors in stock, looking for a pair with the desired ratio:tolerance in some Thevenin range. That has saved me untold hours of misery.

an

te:

loman

gill

So it's a real invention, Have you patented it or published it?

Cool? We all push stuff until breaks - you don't really know how it's going to break until you've managed to break it - but describing this as "cool" is sort of odd.

I don't assume that you are lazy, I assume that you don't know much (incorrectly in this instance, if you went to trouble of finding out about Bracewell's book and getting hold of it) and prefer fiddling at the bench to thinking at your desk.

It's prudent - rather than lazy - to take advantage of known - and proven - techniques that have been published in the literature, rather than trying to invent something new. Sometimes you've got no alternative, but invention is a risky business and can be dangerously time-consuming.

-- Bill Sloman, Nijmegen

he

g.

I did that by hand often enough at Cambridge Instruments that I pretty much knew what was available. I found it quite fun. Misery was preparing project cost and time estimates for my bosses, particularly when they wanted me to re-do them on the basis of extra engineers I knew very well that we were never going to get.

-- Bill Sloman, Nijmegen

I sure didn't invent my deconvolution algorithm by fiddling at a bench. But experimenting with real parts is necessary and productive, too.

Inventing stuff is fun, and not as time-consuming as doing nothing.

And people may more for stuff that nobody else makes. Anybody and his twelve dogs can copy a circuit from your beloved literature.

Patenting only makes selective business sense, usually not. Trade secrets are best, when possible. Publishing trade secrets is folly.

-- Cheers, James Arthur

xes

ax

...

Yup. You can avoid the loss by holding things with actual value, like scrap iron, oil, houses, or even companies. Those scale roughly with the currency, long term. Short-term, not.

It sounds like Spain, Italy, and France are beating us to it. Their borrowing rates are hitting critical, which will then spread throughout Europe.

There's a plan out there called the "penny plan." Cutting spending just 1% a year brings us to balance in just 8 years, and does a bunch of other good stuff.

Of course, we won't do it. That'd be too easy.

-- Cheers, James Arthur