Friday, March 24, 2006

Cause and effect...

Cause and effect.  I was reading this page the other day and found it thoroughly enjoyable.  Bread is dangerous indeed!  False causality, we see it all of the time – it is the cause of many myths – probably the cause of most myths relating to Oracle (single extents are best, reorganizations are always necessary, full scans are evil and the like).

I really liked that example.  Some of the ‘facts’ are really funny, like “More than 98 percent of convicted felons are bread users.”.  That is likely very true, but not relevant.  “More than 90 percent of violent crimes are committed within 24 hours of eating bread”, again true – but related? “Bread has been proven to be addictive. Subjects deprived of bread and given only water to eat begged for bread after as little as two days” – too funny.  

But my personal favorite: “Most American bread eaters are utterly unable to distinguish between significant scientific fact and meaningless statistical babbling.”.  Now that was awesome.

Now, since that was light and fluffy – you can also read this more serious paper.  Cary Millsap forwarded that onto me a while ago.  Thought provoking.


Blogger Andrew Allen said....

I will recommend you the book "The Demon-Haunted World: Science as a Candle in the Dark" by Carl Sagan. I have it as a book on tape and listen to it when I need reminding about how import science (and the scientific approach) is to us. I think it should be required reading for anyone aspiring to teach our children -- especially those who teach science.

Fri Mar 24, 03:12:00 PM EST  

Blogger Andrew Allen said....

If you like the bread post, you will love this one...

Fri Mar 24, 04:06:00 PM EST  

Blogger Andrew Allen said....

Or. . .

Fri Mar 24, 04:08:00 PM EST  

Blogger shrek said....

there's lies, damn lies and statistics.;-)

Fri Mar 24, 04:32:00 PM EST  

Anonymous L Seley said....

On Dr Lamport's web page of his writings he has a paper on "Specifying Concurrent Program Modules". In the abstract it says

"A specification should say precisely what it means for the system to be correct, so that if we prove that the system meets its specification, then we can say that the system really is correct."

(does that scare anyone else as an assumption of correctness?)

If that's his underlying premise then spec designers need logic classes.

There will always be people who "live in a world of homeopathy and faith healing". They're the ones who are so clueless that they think that cleaning the monitor will keep their PCs from freezing (Trying His Best).

Is it logic that guides the developer and helps him decide to use the drag-and-drop widget so users can drop their Outlook attached files onto a program icon? Or is it really be a desire to "Enchant users"? They're not necessarily the same thing.

Fri Mar 24, 05:15:00 PM EST  

Blogger Robert Vollman said....

"I realised that there was something very wrong with the way that system administrator's mind worked."


Fri Mar 24, 06:05:00 PM EST  

Blogger Kurt Graustein said....

The "Future of Computing" paper is certainly thought-provoking. The first thing that came to mind was Toon Koppelaars' presentations at Hotsos Symposium 2006 entitled "Applied Mathematics for Database Practitioners". I know you attended it, Tom. Did you find it as interesting as I did? The ability to represent our data, and the relationships between each entity, using mathematical logic is very powerful.
When asked to assist an application developer with SQL tuning, my first request to them is to explain in "plain English" (my language of choice) exactly what they expect the query to return. I often find that what they want is not what the query is retrieving. Toon's presentation arms me with the language to prove it. Even better, though, it helps me prove that alternative versions of the query, which perform better, are producing results that are logically equivalent to the original query.
The second thing that came to mind from the paper was the concept of design and testing as used in Extreme Programming: building the test before writing the code, breaking a problem down into its simplest forms/functions and tackling the code one small story at a time, etc.

Fri Mar 24, 07:48:00 PM EST  

Blogger Thomas Kyte said....

he "Future of Computing" paper is certainly thought-provoking.

Yes, I found Toons presentation to be one of the high points of the conference

Fri Mar 24, 07:56:00 PM EST  

Blogger Dougie McGibbon said....

"there's lies, damn lies and statistics."

... damn statistics, delivery dates and benchmark tests.

The Devil's DP Dictionary - circa 1981

Sun Mar 26, 02:41:00 PM EST  

Anonymous Barto said....

In the line of the dangerous bread, the infamous Dihydrogen Monoxide:

I also recomend 'The Demon-Haunted...', as I recomend all Sagan's books. I think the active skepticism is the first thing a consultant needs to do his/her work.

As a skeptic, It was a pleasure to read this:

Mon Mar 27, 10:10:00 AM EST  

Blogger Robert said....

It seems ironic that in a paper about logic, the author implies that because faith healing cannot be proven scientifically that (therefore) it is untrue.
(or maybe I am mis-reading/interpreting him?).

Mon Mar 27, 02:27:00 PM EST  

Blogger Thomas Kyte said....

I don't see the author saying anything about proving it scientifically or otherwise. Calls it a superstition

And since the human body is so complicated and not well under-
stood, people have come to believe in a number of medical superstitions like
homeopathy and faith healing.

Mon Mar 27, 02:44:00 PM EST  


<< Home