Saturday, May 02, 2009

Paranoia: it can affect IT testing: production problems can have hidden "contagion"!

All the “hysteria” over possible pandemic flu and the change in attitude toward “presenteeism”, particularly in information technology, makes me ponder another point. We have different experiences with how much “risk” we will tolerate.

In the I.T. workplace, that can translate into “paranoia” that something can go wrong after a major change has been elevated to production. It’s very difficult to anticipate everything that can go wrong in making up a test plan.

In my experience, in a more “conventional” mainframe business IT career, one particular area would be archiving of data or reports, as on microfilm, data warehousing or various backups. Since users didn’t typically need to refer to these, it was possible that some item might not be getting saved when it needs to be, and not be noticed for months or years. I remember a couple of moments of panic, one time running in to the office twice to look at an image copy that no one had checked, another time not being able to find a fiche at all. Back in 1976, we actually did a “tape save” of a company’s accounting data with a COBOL program but never bothered to check that it would be usable.

Self-preservation in those days meant maintaining CYA hard-copies of dumps and tests, often in cabinets. Today, we might see such a practice as a security problem, possibly compromising the privacy of consumers.

The other area where there will be “testing paranoia” would be in Internet deployment of fixes. I can imagine the care that would have to go on at Microsoft before pushing monthly security updates through automatic update feature, or McAfee or Norton before replacing anti-virus engines as well as DAT files.

No comments: