False testing claims

greenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread

I am getting tired of reading reports where testing is reported in a misleading way.

Ex. - what's the difference between these 2 statements:

1. Invent,Assessmt.,Remed., Testing of remediated/repaired systems 2. Testing to look for failures as a part of the assessment phase

Now I'm getting very tired of the reports (especially in the utilitiy sector) who are using testing in case 2. to represent testing in case 1.

I was recently at a y2k information meeting with a regional power producer and distributor and they were doing just that. The recent NERC report is doing the same thing. Since everyone knows that testing is the final step, it makes you appear farther along than you really are.

-- James Chancellor (publicworks1@bluebonnet.net), January 14, 1999

Answers

I just read an article called "How to Detect Y2K Spin" at: www.y2knewswire.com/y2kspin.htm It's a must read because it basically echoes what you have said here.

-- bardou (bardou@baloney.com), January 14, 1999.

Good one bardou! Hotlink:

http:// www.y2knewswire.com/y2kspin.htm



-- Diane J. Squire (sacredspaces@yahoo.com), January 14, 1999.


While reading NERC audit reports, I noticed how the wording was carefully constructed so as to be confusing. (Yet more fodder for debates on forums like here!) It is clear to me however, that if the word "testing" is used in sentences when refering to the assessment phase, it means "look for" "research" "find out" to see if a system is susceptible to y2k failure. If the system is deemed prone to failure, then remediations are done and then move on to "testing phase" of remediated system, to "assertain" that the remediations were done properly. I can't see any other logical ways to do this job. No matter how they report it.

-- Chris (catsy@pond.com), January 14, 1999.

I saw one company, dealing mostly with embedded systems, who broke theirassessment phase into several subphases:

1) Discovery. Find every embedded system, and determine whether any date is used in that system.

2) Evaluation. Test each system that uses a date to determine if there is any aspect of noncompliance.

3) Determination. In each case of noncompliance, decide whether the system can be discarded/disconnected, or whether there is some workaround (like turning the date back), or whether it can be repaired, or must be replaced.

The remediation phase consisted of doing the setbacks, repairs, and replacements.

The testing phase was to consist of returning to the evaluation subphase of assessment for each remediated system, and repeating the process from there until no more noncompliance could be found, one system at a time.

This doesn't strike me as an unreasonable plan, and it permits every step in the process to proceed largely concurrently when there's a large number of devices to deal with. But it does lead to some ambiguity as to which iteration of the evaluation step is being performed on any given system. And since some systems are much more important than others, it renders nearly meaningless any summary of how close they are to completion (or to being effectively functional, which can be a very different beast).

-- Flint (flintc@mindspring.com), January 14, 1999.


Flint, thats why there is usually something like

1.5) Triage Determine those systems that are actually crucial to the company's business, and the rest of the work will be applied to these systems only.

Any Y2K project that I have ever heard of is always basing their facts and figures on these "mission critical" systems. Attempts to somehow explain lack of progress as being due to statistics involving non-critical systems is laughable.

-- Jack (jsprat@eld.net), January 14, 1999.


Moderation questions? read the FAQ