Month: December 2012

Deming’s Seven Deadly Diseases

Posted by

3

William Edwards Deming was sent to Japan in the early 1950’s and propagated his ideas about quality control and production process throughout Japanese industry.

There’s a wealth of wisdom in Deming’s work, albeit much of it industrially focused, but I’m particularly fond of his “Seven Deadly Diseases” of management (with my comments):

  1. Lack of constancy of purpose
    • It’s clear that having some core concepts about what you are trying to do is helpful… simple, effective statements about what’s important to your company, what your company does, and perhaps what your department’s role is in helping to fulfill the company’s purpose.
  2. Emphasis on short-term profits
    • Encourages what Bob Lutz describes as what-can-we-get-away-with thinking.
  3. Evaluation by performance, merit rating, or annual review of performance
    • These systems reward results rather than process-improvement, which can be counter-productive, and thereby encourage workers to maintain the status quo rather than innovate… their goal is to ‘get it done’ rather than to improve how they do so.
  4. Mobility of management
    • Too much ‘reorganization’ interrupts and breaks process improvement initiatives. Probably happens so much because of #3.
  5. Running a company on visible figures alone
    • You cannot measure everything, but must nonetheless do things you think need to be done. Too many times are we told not to do something if you can’t show me it will be valuable.
  6. Excessive medical costs
    • A very interesting observation made over 60 years ago
  7. Excessive costs of warranty, fueled by lawyers who work for contingency fees
    • Maybe not so relevant to the software industry.

It’s also worth mentioning a few other items from “A Lesser Category of Obstacles”:

  1. Neglecting long-range planning.
    • I’m a little torn on this.  In software, too much long-term planning can be a waste of time, but you certainly can’t neglect it entirely.
  2. Relying on technology to solve problems
    • I see this all the time. Figure out your problems first please… I beg you… before you start buying software you think will solve it for you… it won’t.
  3. Seeking examples to follow rather than developing solutions
  4. Excuses, such as “our problems are different”.
  5. Placing blame on workforces who are only responsible for 15% of mistakes where the system designed by management is responsible for 85% of the unintended consequences.
  6. Relying on quality inspection rather than improving product quality.
    • Relying on software testing rather than changing how we build the software in the first place.

I’m often drawn back to these little pearls of wisdom and I continue to be amazed at the prevalence of many of them throughout the industry. Keep them in mind while you are trying to steer your own efforts.

Find stored procedures that won’t compile

Posted by

2

Here’s a handy little code snippet.  Find any stored procedures that won’t compile on SQL Server.

The routine will list all procedures, showing ‘– Failed’ with the error messages for those that fail.

Note that recompiling the procedures may also throw dependency errors also that can’t be caught with try/catch block … these errors likely indicate an issue as well.

SET NOCOUNT ON

DECLARE @ProcedureName VARCHAR(2048)
DECLARE @ProcedureBody VARCHAR(MAX)

DECLARE @RoutineName varchar(500)

DECLARE procCursor CURSOR STATIC FORWARD_ONLY READ_ONLY
 FOR
 SELECT
 --TOP 1
 SCHEMA_NAME(schema_id) + '.' + NAME AS ProcedureName,
 OBJECT_DEFINITION(o.[object_id]) AS ProcedureBody
 FROM sys.objects AS o
 WHERE o.[type] = 'P'
 ORDER BY o.[name]

OPEN procCursor
FETCH NEXT FROM procCursor INTO @ProcedureName, @ProcedureBody

WHILE @@FETCH_STATUS = 0
BEGIN
 -- Might have to play with this logic if you don't have discipline in your create statements
 SET @ProcedureBody = REPLACE(@ProcedureBody, 'CREATE PROCEDURE', 'ALTER PROCEDURE')

 BEGIN TRY
   EXECUTE(@ProcedureBody)
   PRINT @ProcedureName + ' -- Succeeded'
 END TRY
 BEGIN CATCH
   PRINT @ProcedureName + ' -- Failed: ' + ERROR_MESSAGE()
 END CATCH

 FETCH NEXT FROM procCursor INTO @ProcedureName, @ProcedureBody
END

CLOSE procCursor
DEALLOCATE procCursor