MASER's

He has a chapter that discusses that very issue.

But he sure has a lot of references about entire academic and scientific enterprises being very wrong.

This is interesting:

Excel forms are notorious for coding (or whatever you call it) errors.

We have a customer who sends us absurd buggy-Macro Excel froms to fill out. I have several times tripped across the date bug, wherein Excel decides that a numric field is in fact a date. Once it does that, it seems impossible to fix, so we just leave the fields as dates. Nobody seems to actually use the completed forms, so it works out.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

This might reinforce that assertion: I have it on my web pile for instant reference whenever I find myself becoming excessively judgmental. I haven't read the book, but I suspect that the root of the problem is that in order to sound convincing, it is often useful to pass oneself off as an expert. It's not that experts produce wrong conclusions. Rather it is that some people produce wrong or premature conclusions and then try to deflect criticism by elevating their credibility. They do this by acting like an expert, often not in their areas of expertise. To the casual observer, it's difficult to distinguish if it's an expert that has made a mistake, or if it's a mere mortal trying to act like an expert.

I have this problem, which appears to be incurable. Perhaps I should write a book on how to act like an expert?

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

But...but...but... this time it's different!

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

Then that version of Excel is buggy. Try using TEXT for field attribute.

Yes! Put the em-phasis on the concept of "act", and maybe as a court jester..

• I like the quote: "Physics, as we know it, will be over in six months." -- Max Born, 1928 As for timing he was not that far off (A-bomb experiments for one).

• This one is rather accurate: ""That the automobile has reached the limit of its development is suggested by the fact that during the last year no improvements of a radical nature have been introduced." -- Scientific American, 1909 NOTE NO improvements (except for materials used).

Nope. Experts offering incorrect predictions usually don't get a second chance.

The big problem with being an expert in any particular technical field is that the field grows quicker than most experts can keep up. If the area of expertise includes adjacent specialties, the rate of expansion and volume of additional knowledge required to stay up to date can be overwhelming. What this means is last years expert is well behind the times this year. Or more simply, experts tend to have an expiration date. Little wonder the experts tend to be wrong.

Another problem is that if an expert makes one minor mistake, it implies that all of his previous work is now suspect. The Hollywood heroes of my youth mostly failed to age gracefully. But, so did the scientific heroes of my youth fail to keep their reputations above the water line.

If you ask an expert how something works, they will usually not tear it apart, look inside, make measurements, watch the device in action, or do all the things commonly performed by a person of lesser intellect. Instead, the expert will dig into their past understanding of how something similar worked in their past and deduce, possibly from first principles, how they might have designed and built such a device. So, instead of description of how something worked, we get a description of how the expert might have built it. Little wonder that subsequent conclusions are usually erroneous.

If that were all there was to being an expert, no company would ever want or need to hire one. They could do as well with a recent college graduate. So, why do they hire experts? Because experts have made numerous mistakes in their past, while a recent graduate has not. The idea is to not repeat those mistakes, which is really the main value of hiring an expert. After you've done everything possible wrong, whatever remains, no matter how improbable, must be right. Therefore, if being an expert is to remember and prevent their past mistakes, it would not be unreasonable for to expect them to continue to make new mistakes, resulting in the premature judgments listed in the above URL.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
Santa Cruz CA 95060 http://802.11junk.com 
Skype: JeffLiebermann     AE6KS    831-336-2558

So in Rb, optical pumping is all about "pumping" (making a population inversion) in the ground state. There is not one ground state but a whole bunch, separated magnetically, by Zeeman and/ or hyper fine interaction. You then get a signal by stimulating transitions within the ground state.

The details of the pumping process get messy, but you can point to one ground state that has zero (optical) absorption probability. (Because of electric dipole selection rules, you use circularly polarized light, and one state has m_sub_z = N and there is no excited state with N+1. (delta m_sub_z = +1) Decay from the excited state has no such selection rule and some atoms decay into this "dark" ground state, creating the population inversion.

George H.

That's a special case of all versions being buggy. It IS a Microsoft product.

We just get the forms to fill out. If a numeric field wants to promote itself to a date, we're not going to try to fix it.

This particular customer thinks entirely in Excel and PowerPoint. I don't mean "thinks" literally.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

Freedman's point is that the majority of published, peer-reviewed research is not reproducible, and is typically mistaken, and that's how the scientific institution works nowadays. Retractions are very rare, like a cop ratting on another cop. The other type of wrongness is institutional, where the dogma of an entire field of study is accepted, enforced, and wrong. The most profilic wrong-paper-writers are the heroes. They get many second chances, often an entire career of being wrong.

Freedman specifically excludes engineers from his analysis of wrongness. That makes sense: everything that we build is experimentally tested.

Disagree. The big problem is that the experts are in control, and are wrong, and punish dissent. Progress happens when the owners of the accepted paradigm die. They are too powerful to challange in their lifetime, and too invested in being wrong.

The idiots who engineered the last few economic meltdowns are *still* revered, and paid, as experts.

If the

Exactly. But they are still in charge.

Or more simply, experts tend to have an expiration

Retirement or death.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

Yup. The recent meltdown in neuroscience over the statistical significance of their MRI "discoveries" is a classical example.

Another confirmation (if any were needed) that any field with "science" in the name isn't one.

And the quality of papers I get sent to review (optics and optical engineering mostly) has declined a lot over the years.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net

I don't suppose you have that option, but when people send me spreadsheets to fill in, I answer in ordinary flat text. Let them fill in their own damn forms.

Jeroen Belleman

We have a new, somehow mandatory, Excel form to fill out for every unit that we ship, to verify that every one has been visually inspected at about 15 steps in the production process. The goal is to produce "continuous improvement" our DOA rate. Our historical DOA rate with this customer is precisely 0%.

Sometimes when we fill out a box it decides to become a date, not a number. We just let it happen. I'm sure the completed forms go into some database and never come out.

They also use Outlook, which can be amusing.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

FMRI is a poor substitute for witchcraft. I did volunteer for a series of MRIs, after my crash, in a brain-damage study. There will be thousands of participants, so it might have some statistical value.

Freedman is also really down on mouse studies.

The hard sciences at least have experimental verification. That tends to keep things from wandering too far from reality.

Just to have an open mind, I ordered the most popular introductory Sociology text and forced myself to read most of it. Amazing stuff.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

Hmm, don't the socialists, like Slowman, say something similar?

Correcting the field attribute to get the kind of representation that shoul d be there isn't exactly a big deal. Excel is cranky enough that it's fairl y easy for it decide to use the wrong representation, and fixing that kind of fault has to be one of the basic Excel skills.

Obviously not. Excel - or more precisely VisiCalc - was invented to allow n on-programmers to do a kind of programming. I'd still prefer to do stuff in Fortran, if I had the choice.

PowerPoint is useful for presentations, and has more text-fiddling options than Word and Libre Office, but it is rather specialised.

Maybe they dumb down their output so that you can follow it?

--
Bill Sloman, Sydney

Most of it is actually reproducible. The people who run statistics on their data often don't know enough about statistics to run the right statistics or understand exactly what the statistical data is telling them, but the er rors aren't generally serious enough to make their conclusions mistaken.

90% of what gets published is uninspired and not all that useful, but "mist aken" is a gross exaggeration.

You are thinking of climate science, and the problem there isn't with the f ield - which has done brilliant work, so we now know how and why the earth switches between ice ages and interglacials - but with the fact that the re sults of their work are inconvenient to people who make money by digging up fossil carbon and selling it as fuel.

That bunch of greedy crooks has spent a lot of money on rather crude anti-s cience propaganda, which shows up on the kinds of denialist web-site that g ullible suckers like you lap up and propagate as gospel truth.

Nobody in science punishes dissent. Progress happens rather faster than you like to think, and genuinely brilliant new ideas take over fields very rap idly.

Einstein's four famous papers in 1904 made him famous more or less overnigh t. It took a while for the implications to sink in.

His paper on Brownian motion finally confirmed the physical reality of atom s. Ernst Mach didn't die until 1916, and remained sceptical about the idea until he died, but it didn't hold back the rest of the field.

You mean the Chicago School? They produce the kind of nonsense that the tax

-evading rich - and James Arthur - like to hear. They aren't any kind of sc ientists, but rather a branch of the entertainment industry, producing soot hing nonsense for the very rich.

The neo-Keynesians who get us out of the economic melt-downs that the Chica go school contrive to engineer are the actual experts, and rightly dominate the field.

Experts do try and stay on top of their field. My own experience is that wh enever I get faced with an old problem, there are new solutions out there t hat I have to get my head around before I can come up with a more or less c urrently optimal solution.

My 1996 millidegree temperature controller paper was a contribution to a ve ry well established field - I'd even made a minor contribution myself, back in 1978 - but by 1993 sigma-delta converter chips were around and fairly c heap, so I could describe a new way of solving an old problem.

--
Bill Sloman, Sydney

Not exactly. If you've got a a couple of million dollars worth of MRI scann er in your basement, you are under considerable pressure to keep it busy an d churning out more-or-less publishable research.

The machines generate monumental amounts of data, and neuroscientists aren' t actually trained in data-mining, so there' a obvious risk that they'll la tch onto spurious correlations.

Neuroscience hasn't actually "melted down", but there is a much wider perce ption than there used to be that you need to run your proposed statistical analysis past a statistician (my cousin the statistician is formally retire d, but I'd be happy to pass on his e-mail address) before you start spendin g money, and expensive machine time, on collecting data.

Perhaps an over-generalisation. Some of the people involved in neuroscience were well aware of the problems.

When the Donders Centre was set up in Nijmegen in 2000

the academics in my hockey team went in for a certain amount of malicious g ossip about the problems.

The Donders Centre machines were not in a hospital, so any medically qualif ied investigators involved were constrained by people who understood scienc e and statistics. MRI machines in hospitals are more likely to be exploited by publication-hungry medicos who know little about the scientific method and less about statistics.

Or your standards have gone up. We all get testier as we get older.

--
Bill Sloman, Sydney

FMRI isn't remotely like witchcraft. It provides real information about wha t's going on inside your head in a way that's otherwise impossible. It's ra ther broad-brush information - there are some 86 billion neurones in there and the resolution doesn't get anywhere near down to neurone or synapse lev el - but it can answer some questions.

That kind of study needs to see a range of brains, including those that don 't work all that hard.

So he wants everything tested on representative samples of humans? Mice are a lot cheaper, and close enough for many purposes.

There's enough individual difference between humans to create problems, and the pharmaceutical industry habit of using young white male volunteeers fo r clinical trials has occasionally produced misleading results.

Which one? Any study with Diederik Stapel's name on it is suspect.

But physics had N-rays, albeit a century ago. Stapel was a conscious cheat, but the physicists who thought that they were seeing N-rays were deceiving themselves.

--
Bill Sloman, Sydney