Tuesday, May 10, 2005


In your article "In search of the truth Or Correlation is not Causation" you mentioned:
"The other one of my guiding principles is that I ask for proof from everyone (yes, it can be annoying, it carries over into my personal life as well. Ask my wife and kids)."
It would be interesting if you can expand on affect of your approach to personal life. In today's interconnected world, most of us are specialized and need to trust other experts (doctors, financial advisors, car mechanic etc.) on their word.

Well, it certainly does spill over. Take doctors for example. (We can pretend for a moment that the developer is instead a physician and the application is the patient). There are many types of MDs

  • The emergency room (ER) doctor These physicians do "triage," separating the hopeless patients from those that can be helped, performing quick-fixes along the way to keep patients alive as long as possible. They will take a heart-attack patient with a history of smoking, bad diet, and no exercise and stabilize their condition.
  • The operating room (OR) doctor The OR physician gets the patient after they've gone through triage and after the ER doctor has temporarily patched them up. The OR doctor strives for long-term fixes to keep the patient not only alive but functioning as well as possible. They perform the by-pass operation on the heart attack victim, attempting to clear the arteries.
  • The physical therapist (PT) The PT gets the patient when the OR doctor is through and begins the long and painful (not to mention expensive) process of rehabilitation.
  • The preventative medicine doctor These physicians strive to avoid the preceding three doctors at all costs. They counsel the patient to quit smoking, eat a healthy diet, and exercise, developing a multistep program to get them in shape. If they do their job right, barring unfortunate incidents (like a car accident), the patient will never see the ER, OR, or PT doctors.

In the above list, I would expect to have to "trust" to some degree the ER doctor. It is fast and furious and they are making quick decisions based on experience and limited facts. But the decisions must be made quickly (or they may never be made at all). They do not make big changes - their goal is to stabilize and get the patient ready to be fixed. And -- we try to avoid the situations that necessitate the ER doctor all of the time (don't we?). For example, I avoid trimming my hedges with my lawn mower - it would work but would really increase the change of my needing the ER doctor wouldn't it?

So that ER doctor, we must to some degree trust, but they are doing quick patches, quick fixes. We don't want to use them day to day - to figure out "how to live properly", and we avoid them at all costs. That we need to use them indicates something went wrong in the first place.

The other doctors however, are different. We demand test and proof from them don't we? Before I get an operation, I'll get multiple opinions - I'll have reviewed the test results - I might even have studied the problem myself to some degree. I'll have asked for information from multiple people over time. We would have tested this, we would have gotten some evidence on this. If it was an experimental treatment, we would go into that with eyes wide open. Same with the other doctors. If I was going into physical therapy, I'd want some evidence that what they were doing was the best way to fix what ails me. We'd have test, evidence, facts to see past successes and failures. We'd have a pretty good idea that what we were going to undergo would actually be successful and the degree to which it would be.

Sure, there is a matter of trust there, but really - it is all backed up ultimately by facts, by evidence. Before we get surgery, we have tests performed. We would never let a surgeon just poke around and do stuff because it fixed another patient in the past right?

I think that applies to everything. I tend to research things - when I went to buy my hybrid gas/electric car - way before I went near a dealership, I learned quite a bit about them. I read the opinions of people. I read the technical articles. I read the good and the bad. I knew the dealer would say "50-60 miles/gallon", but I knew the reality would be 40-50 (and it is, 45-47 miles/gallon is my year long average). I understood how they worked (but don't ask me to work on one - I opened the hood and saw not much that I recognized). I understood what the car would and would not do for me. The dealership only had to let me drive one to fill in that last piece of missing information -- would I find the car comfortable to drive.

Sometimes we don't have the time/experience to get in depth in some topic and there we rely on "experts in the field". Take wine for example - I like to drink it (any flavor as long as it is red). I don't know very much about it however. I go to a vintner (we happen to have one close by). I trust him to pick out a case of it for me. I do this because he records what I've had in the past - and remembers what I liked and is able to now make recommendations for me based on past experience. But what he learned about me applies to me and me alone (see correlations to computer systems?) It would not apply to you (maybe it would, but probably not). It is like training an expert system - it needs some feedback to refine the suggestions. The wine we get now is usually pretty much dead on - it has been a long time since we've had some we didn't like. The vintner took our price sensitivity, our personal tastes and based on their past experience - is able to suggest something for us we'll probably be OK with.

But I would not walk in there and let someone else pick it out for me - not without expecting some wine we aren't going to like so much. Even if that person was expert - because we don't have that history yet. I'd have to train them (continuity is a good thing as well - many correlations to computer systems) all over again.

So yes, I'd say I have to either have lots of evidence, or for some things have the knowledge that the person dispensing the advice actually knows my unique circumstances. Maybe that is why I'm down on ROT (rules of thumb). It is not that by finding a single exception they are ruled out - but rather that literally thousands of exceptions exist for each one - for each rule of thumb - and if you try to apply a rule of thumb that has worked for some one else, without understanding "why" you'll do one of three things:
  1. make things better
  2. make things not change at all (but have expended a lot of energy in doing so, and perhaps set expectations high)
  3. make things much much worse

The vintner is full of rules of thumb, but until they get to know me - they are just guessing.


Blogger melanie caffrey said....

Very well said.

I've reviewed "software solutions" proposals in the past from so-called "impartial" VARs.

I've often found that

A) "Impartial" is an arbitrary term, depending on who you're dealing with, and
B) I never trust those proposals that name a particular technology and tell me it won't suit my needs because it is "inefficient" or "It doesn't work as well as "XYZ" ", and
C) I tend more to trust those proposals which state something along the lines of "We haven't personally worked with XYZ, but we think it worthwhile to run some comparative tests against ABC, and see which one you think might best suit your needs."

And of course, before any such comparative tests have been done against XYZ and ABC by the VARs (:) ), on the side I will have done LOTS of reading up and (time and money allowed) playing with XYZ and ABC. (And others, as well ... )

... similar to your hydro/electric car anecdote.

Why leave a big, expensive, possibly life-changing decision up to somebody else implicitly?

No matter how much experts may know, they're not usually the ones who are ultimately stuck with the results of their advice.

Tue May 10, 11:02:00 AM EDT  

Blogger David Aldridge said....

One of the most valuable classes I took as part of my engineering course was Policy Analysis, and I came away from it with a healthy scepticism about all opinions, allegedly impartial or otherwise.

It's now automatic that on hearing or reading an opinion I consider "what would this person gain from me believing this?", and also "what are they not saying here?"

Particularly vauable when someone is telling you to just trust them based on their "status" or on the number of books that they've self-published, for example ;)

(I don't know why we are supposed to be impressed by the number of books that a person has published, when they control the publishing company!)

Tue May 10, 11:17:00 AM EDT  

Blogger Rachel said....

I know of very few people who just assume that the expert is right when it comes to things in their personal life. Insurance companies encourage "second opinions", you get several quotes when doing construction, you even comparison shop prices on computer equipment (heck, I comparison shop prices on groceries!).

So yes, I think we tend NOT to just rely on the experts in our personal lives... why do so many people do so in their professional ones?

Tue May 10, 11:26:00 AM EDT  

Anonymous Anonymous said....

Hi Tom,

that can lead to a philosophical question : can you rely on a single fact without proving it by yourself ? Ad adsurdum you couldn't even trust Einstein without doing the same calculations ;-)

Tue May 10, 11:33:00 AM EDT  

Blogger Thomas Kyte said....

that can lead to a philosophical question

but hopefully, this article showcased that I rely on these things all of the time.

When I bought my car, I went into the deal with my eyes wide open -- I believed that the miles/gallon would be about 40-50, not 50-60. Why? Not because I got on and performed the test, but because many people that owned the car were willing to share their experiences and let us know what they actually saw. (side note: the warmer the climate, the better the mileage with the hybrid -- Canada won't do as well as Southern Florida -- I am in a mixed zone, right in the middle).

I trust Einsteins theories because many people in the field (a field I am decidely not expert in) have bashed on them hard -- they are looking for ways to refine/break them (for if you find a mistake with Einstein's work and can prove it to the satisfaction of others -- might that have some effect on your own career). So, I trust that those theories are the conventionally accepted truth (for now, Newton had his day).

I rely on facts without running the test cases myself all of the time. The fact that others in the field have taken the test cases, used them, expanded on them, looked for ways to break them -- gives me the confidence there. But, in my field -- that of the database -- I run the test cases all of the time. My inbox is full of them.

So, I don't need to reproduce Einsteins work -- for it has been done many times for us (and would take a massive conspiracy on a world wide level to be fabrication at this point)...

Tue May 10, 11:59:00 AM EDT  

Blogger Jeff Hunter said....

I ask for proof from everyone (yes, it can be annoying, it carries over into my personal life as well. Ask my wife and kids).
Tom's dinner table: "Please provide proof and a cost benefit analysis of why you need that new bike, son."

Tue May 10, 12:19:00 PM EDT  

Anonymous Chris Gait said....

I liked the medical analogy. I find a lot in common between database work and medical diagnosis and treatment. One of my favorite aphorisms for database optimizations comes from medicine: 'When you hear hoofbeats, think horses, not zebras.' I have to read that to myself often, like a mantra, to avoid looking for obscure parameter settings and anomalies while missing that the client is simply missing an index on an important column or the like.

Tue May 10, 12:27:00 PM EDT  

Anonymous Anonymous said....

> I rely on facts without running the
> test cases myself all of the time.
> The fact that others in the field
> have taken the test cases, used
> them, expanded on them, looked for
> ways to break them -- gives me the
> confidence there.

I think one of the signs of the true, deep professional is the ability to stand back, look at the overall situation (_not_ the "big picture"), and say, 'hmmm - this is not going the way we expect. Perhaps it is time to start questioning some of the accepted facts'.

Or to put it another way: we have all heard the "when you hear hoofbeats think horses not zebras" saying, and I do use it daily. But I also lived down the street from a zoo once, and about twice a year I would have to call the police to come collect some exotic animal from my backyard (usually birds, never a real zebra - but a springbok once). So I also try to keep in mind the thought there there just might be a zebra out there.


Tue May 10, 12:46:00 PM EDT  

Blogger Thomas Kyte said....

Tom's dinner table

A bike -- I wish, last night it was in fact a webcam that can follow movement with a built in microphone -- so he can chat with his friends...

Tue May 10, 12:50:00 PM EDT  

Blogger Bundit said....

Hi Tom,

Not relating to this title at all.

I'm just curious about your published book. Dominando Oracle: Programação Avançada

Tue May 10, 01:32:00 PM EDT  

Blogger Thomas Kyte said....

Not relating to this title at all.

It is a "sellers" page -- hosted on amazon but not really amazon:

...Product Details

* Paperback
* ISBN: 8573931868
* Amazon.com Sales Rank:
(Publishers and authors: improve your sales)

* This page was created by a seller.


it looks like someone trying to sell the Portuguese translation of my book for big bucks

Comments: All books are brand new, Portuguese language. Books shipped direct from Brazil with delivery in 2-3 weeks. Please contact us with special requests.
Sign in to turn on 1-Click ordering.

Tue May 10, 01:39:00 PM EDT  

Anonymous Anonymous said....

Great analogy with the doctors. I found myself nodding the whole time.

Curious, wouldnt this sort of nature label you a "cynic"? Is that good or bad? The word usually has bad connotations most of the time.

Tue May 10, 01:50:00 PM EDT  

Blogger Thomas Kyte said....

wouldnt this sort of nature label you a "cynic"?

Not a cynic, a sceptic would be appropriate.

Cynics are negative, I'm just a doubting Thomas.

Tue May 10, 02:03:00 PM EDT  

Anonymous Anonymous said....

So Tom , Were you trained to be a sceptic or something turned you into one or you were born a sceptic:)

Tue May 10, 02:18:00 PM EDT  

Blogger Thomas Kyte said....

I believe it is genetic, the sceptic.

I secretly desire to live in Missouri.

Tue May 10, 02:36:00 PM EDT  

Blogger Niall said....

Tom's dinner table: "Please provide proof and a cost benefit analysis of why you need that new bike, son."

I vividly remember my Dad when I was about 12 and after we had got bikes for Christmas and the family had got a new TV, saying to me (on 12th Night) "So what's it to be the TV or the bike?" and expecting me to give good reasons for my choice. Man that was hard (and 25 years later the fact that I didn't realize he was pulling my leg still hurts)

I know of very few people who just assume that the expert is right when it comes to things in their personal life

Come to England, we trust doctors, bankers and so on. They're reputable middle class people after all :)

Tue May 10, 04:05:00 PM EDT  

Anonymous Anonymous said....

> tkyte
> I secretly desire to live in
> Missouri.

You're hired - when can you start?


Tue May 10, 04:08:00 PM EDT  

Blogger jimk said....

Cynics are negative, I'm just a doubting Thomas.


Tue May 10, 05:47:00 PM EDT  

Anonymous Paul said....

Come to England, we trust doctors, bankers and so on. They're reputable middle class people after all :)

Doctors and bankers? Surely you've missed out the two most reputable types we've got over here - IT consultants and solicitors (lawyers for you Tom).

Wed May 11, 03:57:00 AM EDT  

Blogger R Menon said....

We demand test and proof from them don't we? Before I get an operation, I'll get multiple opinions - I'll have reviewed the test results - I might even have studied the problem myself to some degree. I'll have asked for information from multiple people over time.

Well said!
So in my scenario (where the car mechanic apparently was being untruthful), you would have already learnt what CV boots (or whatever) are but how would you have known that what he is saying is true? That there indeed is an issue in the car and he is not making it up?
When I first went to him, I read up on his service, asked around. I guess one way could be to look up his service rating in a site (I, in fact thought there is no such service that rates services and thought this could be a good business model - only to discover yahoo does it - though not very prominently:))

In the case of Oracle, you being the expert, have a way of testing even what Jonathan says and I am sure you would have corrected him (or he would have corrected you) some time? Truth be said, after being "affected" by asktom, the idea of proof and the fact that "experts" can be wrong, many times deliberately - did affect my life! Mostly, only in terms of more headache though!:) It is not easy (at least for most people) to form an "informed" opinion on all important topics some of which are as complex or even more so (taxes, cars, a critical operation on a dear one and so on - well those are complex IMHO!:)). I listen to public radio. They usually have discussions on controversial topics and typically announce the expertise of the person as " I would like to welcome Mr. So-and-so, the director of so-and-so, the vice chancelor of so-and-so and so on. They will have multiple experts with contradictory opinions. I always wonder if one of them is not being rigorous enough and I have mostly no time or expertise to . If you look in the field of Oracle, there are only a few, handful of experts who are rigorous - rest are either mediocre or even plain misleading!
I am just pointing out that ultimately the "trust" can, and will in many cases be misplaced - with serious consequences many times!

Speaking of public radio, does anyone listen to "Car Talk"? Here is an opinion on hybrid cars I just found http://www.cartalk.com/content/columns/latest.html

Wed May 11, 09:32:00 AM EDT  

Anonymous Anonymous said....

One of the biggest problem with relying on experts in a field is that they are also fans and often fanatical about their discipline. It's fairly common to see people who do one technology argue with someone who does another technology by reciting documentation to each other as if it's their own original idea.

Most technology specialists lack original thought and recite what other people said or the documentation as a reason to do something. They will twist an architecture around to use their given skillsets. They will constantly talk about how wonderful the technology is that they are specialists in.

I never trust people who recite instead of thinking through an argument. I routinely hear people say you are supposed to do it this way. If you dig deeper it's because the documentation from the vendor says so. They don't dig deeper and try to apply the technology (or other technologies) to the situation and restrictions at hand.

It's also very common for people who are good with different technologies to argue heatedly with each other and all they throw out at each other is what the documentation or other people say. They don't know anything about the other technology yet they act like experts.

Microsoft people insist that what they do is cheaper and quicker to develop with even though they don't know anything about java or oracle.

Java people argue for the cost savings of database independence and to use databases for 'what they are good for' which is storing data. How do they know what a databas is good for?

Sybase people argue that it's the best database because it's built on sound theoretical principles.
Yet I have not found a single one who can recite me any of the Codd rules. If they don't know the theory how do they know Sybase is based on sound theoretical principles?

Oracle people routinely state that you cannot build anything big with SQL Server(some of the largest databases in the world use SQL Server) and often recite you Tom without critically thinking about what you said. That is not to say that you are wrong, it's that I don't trust people who recite.

Ryan Gaffuri

Wed May 11, 10:53:00 AM EDT  

Blogger Thomas Kyte said....

R Menon said....
Well said!

I use Jonathan as a reviewer on my books -- he is an excellent critic. I use many people of his caliber to perform that job.

I've pointed things out to Jonathan, he's pointed out many things to me -- constantly. It is why we share the assumptions and methods.

Wed May 11, 11:55:00 AM EDT  

Blogger Joel Garry said....

does anyone listen to "Car Talk"?

Google about and you can find a number of articles saying you can hose those Hondas out. Interesting meme propagation of a myth.

The Dr. analogy is interesting, Lawson uses it in his book and Daniel Morgan proposes similar licensing for DBA's. I generally agree with both of them. However, the breakdowns in the medical Dr. method are informative:

Most people are in the hospital (I am US-centric, of course) for iatrogenic reasons. Some people who should be in the hospital are not for economic/political reasons.

Not all medical orgs follow the functional organization Tom outlined. My HMO, for example, uses a primary care physician for sorting out which specialist you can see. I had a bad experience once where she didn't think to baseline a blood enzyme, so when the drug she prescribed almost killed me, it was impossible to know whether it was prexisting or due to the drug. (my experiences with the more usual med insurance have been worse). The industry norm is to minimize contact with doctors and maximize profit, not to maximize health. While doctors making snap decisions are right not to think of zebras, they would do well to learn Tom's methodology. They do indeed do things just because it worked on other patients. And they also have stories about people who "researched" things on the internet and have some very strange ideas.

Oracle people routinely state that you cannot build anything big with SQL Server(some of the largest databases in the world use SQL Server)

I have spent the last several days wrassling with a vendor product that works fine with tiny mssql db's, but I'm totally having to hack to deal with my middlin' size Oracle db. Largeness is trivial, what sqlserver (and db-neutral products) can't deal properly with is complex+large+multiuser+transactional.

"So what's it to be the TV or the bike?"

"When I was your age, we couldn't afford a bike" was what I got.

I go to a vintner (we happen to have one close by). I trust him to pick out a case of it for me.


Wed May 11, 08:28:00 PM EDT  

Blogger Kalita said....

No blog for last two days. Busy with the book?

Thu May 12, 10:53:00 AM EDT  

Anonymous Anonymous said....


Trust in a Medical Setting. Hauppauge, NY: Novinka Books, Nova Science Publishers, 2006.

Experience dealing with a host of difficult to impossible situations may help others in their encounters with these difficult and distrusting patients. These individuals may make up a small per cent of patients and family members, probably less than 2 per cent, but take up 90 per cent of energy in coping with day-to-day conflicts that arise from their behavior. Difficulties managing distrustful patients and family members must be dealt with on the spot, and they don’t go away.
Examples come from office experiences or wards, including situations that keep doctors and nurses and therapists awake at night, aggravate waking hours and poison leisure, that is, empirical, based upon experience and observation alone without science or theory. To survive an outrageous patient or relative requires resourcefulness, patience and imagination. Street wisdom learned the hard way is what I present, and without a guide or mentor to soften the bewilderment and sense of failure and frustration that accompanies these individuals. We seldom talk about these difficult, distrustful and sometimes threatening individuals amongst ourselves; rather we suffer and endure them silently, by ourselves. The problem is timeless as recorded in the world’s literature.
Out of the wreckage of human behavior comes valued experience leading to maneuvers and tactics of survival that are appropriate to almost all aspects and settings of human interaction including day-to-day medical care.


Wed May 09, 11:32:00 PM EDT  


<< Home