Click here to monitor SSC
SQLServerCentral is supported by Red Gate Software Ltd.
 
Log in  ::  Register  ::  Not logged in
 
 
 
        
Home       Members    Calendar    Who's On


Add to briefcase 123»»»

Excuses Execrating Excel Expand / Collapse
Author
Message
Posted Saturday, April 20, 2013 12:34 PM
Forum Newbie

Forum NewbieForum NewbieForum NewbieForum NewbieForum NewbieForum NewbieForum NewbieForum Newbie

Group: Administrators
Last Login: Today @ 3:50 AM
Points: 9, Visits: 1,779
Comments posted to this topic are about the item Excuses Execrating Excel
Post #1444733
Posted Saturday, April 20, 2013 2:52 PM


Mr or Mrs. 500

Mr or Mrs. 500Mr or Mrs. 500Mr or Mrs. 500Mr or Mrs. 500Mr or Mrs. 500Mr or Mrs. 500Mr or Mrs. 500Mr or Mrs. 500

Group: General Forum Members
Last Login: Sunday, April 20, 2014 11:44 AM
Points: 561, Visits: 2,417
Excellent editorial. Made me laugh.


Best wishes,

Phil Factor
Simple Talk
Post #1444745
Posted Saturday, April 20, 2013 6:41 PM
SSC Rookie

SSC RookieSSC RookieSSC RookieSSC RookieSSC RookieSSC RookieSSC RookieSSC Rookie

Group: General Forum Members
Last Login: Tuesday, March 11, 2014 8:22 AM
Points: 35, Visits: 69
There is a saying, "A bad workman always blames his tools".
It occurs to me that Excel is the wrong tool for the job in so many cases. The bottom line (pardon the pun) is that the businesses should be getting the right analysis tools instead of blindly relying on a one-size-fits-all solution.
Post #1444773
Posted Saturday, April 20, 2013 7:44 PM
Forum Newbie

Forum NewbieForum NewbieForum NewbieForum NewbieForum NewbieForum NewbieForum NewbieForum Newbie

Group: General Forum Members
Last Login: Monday, April 22, 2013 7:15 PM
Points: 1, Visits: 6
Thank you, thank you, thank you.

When I read reports like this in the newspaper (yes, we still have one of those), I rail against them. My wife can't understand why I am so excited.

Now, if we could just get people to talk rationally about evolution....
Post #1444774
Posted Saturday, April 20, 2013 11:47 PM


Right there with Babe

Right there with BabeRight there with BabeRight there with BabeRight there with BabeRight there with BabeRight there with BabeRight there with BabeRight there with Babe

Group: General Forum Members
Last Login: Friday, April 18, 2014 4:23 PM
Points: 778, Visits: 1,505
Same thing happens with SQL Server (or any other software for that matter). In 2005, in a dysfunctional government department not too far from here, an employee (let's call him Bob) was charged with the task of porting a mountain of SPSS code to SQL Server. The decision was made by his boss, an angry ignoramus who understood neither SPSS or T-SQL. And understood even less, how to deal collaboratively with subordinates.

Bob didn't know much T-SQL himself and thus didn't want to be porting anything anywhere. Bob's boss helpfully decreed that Bob either did what he was told or considered other employment arrangements. So Bob dutifully did what he was told, but with an overriding philosophy in the vein of "If it compiles, ship it." Several errors were identified during the process and invariably Bob would say "It's SQL Server that's the problem - it won't reliably produce the right answers, and talk about SLLLOOOOWWW. Don't get me started on that!" Then Bob resigned, leaving dozens of totally uncommented stored procedures replete with invisible "SQL Server" errors (the worst kind). A classic example was the good old "Let's divide an integer by an integer and then wonder why it produces an unexpected result" trick.
print 1/2
print 1./2
print 1/2.
print 1./2.

That was 2005 and to this day errors still bubble to the surface. The boss's rationale for years of avoiding a comprehensive code audit seems to be
1) If problems are uncovered, it's going to be embarrassing, and
2) The code produces answers that seem about right, so why go hunting for problems that might not even exist.

A lot of sad things came out of this. A staff member resigned, leaving a mess. The acceptance of SQL Server BI (which were far more appropriate tools than SPSS in this situation) in related government departments was set back years because of the misinformation. Saddest of all though was that big decisions affecting some of the most vulnerable segments of the community have been made based on calculations that are probably inaccurate enough to affect those decisions.




One of the symptoms of an approaching nervous breakdown is the belief that one's work is terribly important.
Bertrand Russell
Post #1444776
Posted Sunday, April 21, 2013 7:11 AM


Old Hand

Old HandOld HandOld HandOld HandOld HandOld HandOld HandOld Hand

Group: General Forum Members
Last Login: Yesterday @ 9:45 AM
Points: 310, Visits: 867
Doesn't matter what software you use - if you don't understand your data, or if you ask the wrong questions, you're going to get worthless results. Over and over I tell people, just because a computer spit it out, doesn't mean it's good. Sometimes they listen.

Years ago I had a girlfriend in a PhD. program, and she came to me because her calculator was broken. She was trying to figure out something called Gibbs free energy in a chemical reaction, and the calculator kept saying 'Overflow', so there was clearly something wrong with it. Turns out the formula she was using contained 10^10^6, naturally, far beyong the poor calulator's range. But she had no clue that the formula she was using was nonsense, and if the calculator had managed to spit out the answer in some tidy notation, I'm sure she would have just copied it down and turned it in, with no clue that the amount of energy cited would be far in excess of the total known energy in the entire universe.

I'll go even further - if you don't understand your data, you have no business asking questions of it.
Post #1444793
Posted Sunday, April 21, 2013 9:24 AM
SSC Rookie

SSC RookieSSC RookieSSC RookieSSC RookieSSC RookieSSC RookieSSC RookieSSC Rookie

Group: General Forum Members
Last Login: Tuesday, March 11, 2014 8:22 AM
Points: 35, Visits: 69
I wonder how often software is validated before it is used. Taking Excel as an example, it's a fantastic piece of software and is extremely useful, I use it a lot, but has anyone outside the Microsoft campus taken the trouble to check that the results of various functions are actually correct to the degree of accuracy required by the specific application. When a complex series of calculations are concatenated, how can the compound errors affect the end result? When banks are applying their rocket science calculations using such a tool, how confident can they be that they are getting even close to the correct result?
Now what was that about Sarbanes Oxley and best endeavours?
Post #1444801
Posted Sunday, April 21, 2013 9:03 PM


SSC-Enthusiastic

SSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-Enthusiastic

Group: General Forum Members
Last Login: Yesterday @ 10:48 PM
Points: 177, Visits: 531
@GPO, at least you made me feel a little better about my current situation after that story. It makes me wonder, how do you make an ignoramus like that responsible? Because he probably got a fat bonus for all the 'savings from SPSS licencing costs', while downstream, the public services real customer is suffering.

It's bad enough when the business creates Feral Apps and they become business critical and 'must be supported by default'. It's worse when Feral apps are being built inside IM and people being rewarded for it.

I find myself being dragged into sneaky politics in order to try and throw some light on the dimwitted decisions being made.
Post #1444842
Posted Monday, April 22, 2013 2:39 AM
Grasshopper

GrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopperGrasshopper

Group: General Forum Members
Last Login: Wednesday, March 05, 2014 3:40 AM
Points: 24, Visits: 266
There's a saying:
"Don't blame the tool, blame the fool behind the tool".
Post #1444899
Posted Monday, April 22, 2013 3:03 AM


SSC-Enthusiastic

SSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-EnthusiasticSSC-Enthusiastic

Group: General Forum Members
Last Login: Yesterday @ 8:15 AM
Points: 159, Visits: 797
herbaltea001-winter (4/21/2013)
I wonder how often software is validated before it is used. Taking Excel as an example, it's a fantastic piece of software and is extremely useful, I use it a lot, but has anyone outside the Microsoft campus taken the trouble to check that the results of various functions are actually correct to the degree of accuracy required by the specific application. When a complex series of calculations are concatenated, how can the compound errors affect the end result? When banks are applying their rocket science calculations using such a tool, how confident can they be that they are getting even close to the correct result?
Now what was that about Sarbanes Oxley and best endeavours?


There is some work out there on how accurate the various spreadsheet programs are...
http://www.jstatsoft.org/v34/i04/paper



Excel is an all-round tool, but proper statistical analyses should be done in sas/r/spss/AN Other. I love excel but there is the mistaken belief that because it is on your desktop and anyone can type in it, it is easy - this gives people a false sense of confidence that leads to insufficient thinking and testing.
Post #1444902
« Prev Topic | Next Topic »

Add to briefcase 123»»»

Permissions Expand / Collapse