wird es sich vielleicht richtig auffinden lassen. Bei meinem letzten Vortrag vor 10 Tagen ging es "hochmathematisch" zu, doch der Buchbinder findet sich auch hier im Kapitel 2.2.

Der Vortrag:
What Programmers can learn from Mathematics (applied mathematics)
Hans Hartmann
...
Abstract
There exists a common misconception about the presence of software bugs in today's programs. Commonly it is believed that software failures - in whatever way they are introduced into software - are a result of human errors. Some mathematical results found in the 20th century show that this assumption is not correct. The presentation shows that the very much neglected discipline "program testing" has to be assessed using a different viewpoint when one considers the results of Gödel, Turing or Chaitin. These considerations influence a new way to teach software testers and will show ways to improve on software test management. The presentation is based on several very successful experiences during the past 12 years and will also cover the pitfalls that need to be avoided.

Das angesprochene Kapitel

2.2 The Changed Mindset

Testers will soon discover that they need to form a comprehensive view of the test object. Indeed, they are not testing to ensure that given input fields can indeed be filled in, but instead to confirm that the program runs in a manner that truly fulfills the business needs. Inevitably, this leads us to the question of what the tester is to do in the event that there are no written business processes (in our experience, a case that is frequently encountered). Our answer: You must define the business case yourself. Usually, the client will determine what absolutely must be tested, in the form of a rather casually written business process; and it is the tester’s job to objectify this process. The business process itself represents a test, as one needs to confirm that a coherent process chain exists. Asking the client’s business domain specialists may sometimes resemble running a gauntlet, or the tester may feel like the bookbinder Wanninger (K. Valentin)(6), who gets passed around from one department to the other without finding an answer. As a general rule, however, the domain specialists will be friendly, since they have the opportunity to discuss something that they know very well. Obviously, not even a domain specialist will be able to describe a complete complex business process in the required amount of detail. Thus, it would naturally be much more efficient to have a comprehensive handbook complete with instructions and sample cases. Unfortunately, such a handbook is rarely completed by the time a tester must finish work on the program.
As a result, we make it clear to all of our testers that their task is to find errors, and not to execute test cases. Though test case execution is a necessary evil, we can try to “keep it at arm’s length” by using test automation.

Die erklärende Fussnote:

(6) “Buchbinder Wanninger” is a sketch performed by the German comedian Karl Valentin. In it, the bookbinder Wanninger phones his client in an attempt to find out whether or not he should send the invoice along with the shipment of books that the client ordered. He remains unsuccessful, as he is simply passed from one department of the client company to the next without receiving the desired information. The sketch ends with a completely distraught Wanninger snarling: “Stupid bunch of no-goods!”

Zum Valentin sag ich jetzt gar nichts, weil der steht eh für sich.
Man tät halt so einen noch irgendwie herbei wünschen, quasi jetznan, aber da tät sich der sicher die Zähne ausbeißen an das was eine Realität ist.

Geh bitte, no a Moß auf´m Fallentin.

Der große bajuwarische Sprachakrobat kann eigentlich ohne seine langjährige Bühnenpartnerin Liesl Karlstadt[die heuer übrigens ihren 120ten Geburtstag begeht und mit ihm über 400 Sketche aufführte] nicht bedacht werden, lieber Josef ;-)

Unvergesslich ist mir der "Semmelnknödeln"-Dialog, anläßlich des Wiegenfestes zur Erinnerung ein kleiner Auszug:

Karlstadt: ....man sagt schon von jeher Semmelknödel. Valentin: Ja, zu einem – aber zu mehreren Semmelknödel sagt man Semmelnknödeln.
Karlstadt: Aber wie tät man denn zu einem Dutzend Semmelknödel sagen?
Valentin: Auch Semmelnknödeln – Semmel ist die Einzahl, das mußt Ihnen merken, und Semmeln ist die Mehrzahl, das sind also mehrere einzelne zusammen. Die Semmelnknödeln werden aus Semmeln gemacht, also aus mehreren Semmeln; du kannst nie aus einer Semmel Semmelnknödeln machen...
Valentin: ...solang die Semmelnknödeln aus mehreren Semmeln gemacht werden, sagt man unerbitterlich: Semmelnknödeln.