Tuesday, January 08, 2008

Percipient Subject and Software Quality

Jerry Weinberg in his famous book “An Introduction to General Systems Thinking” mentions quoting Albert Einstein…

“Belief in external world independent of percipient subject is foundation of all science”

Let me extend this to Quality (or Software Quality) …

“Belief in external world independent of percipient subject is essential for the notion of Quality in software world”

All definitions of Quality (except that of Jerry) from Juran, Deming and others somehow assumed the existence of “external world independent of percipient subject”. Hence Quality has been more or less defined in absolute terms.

Let me remind you -
- Quality is value to some person (one who matters)
- Quality is not an attribute of a *thing” – we can not measure Quality but can only assess it in “relative terms”
- Quality is in beholder’s perceptions.
- Quality is not an intrinsic attribute of any “thing”
(Quotes from Jerry, Michael Bolton and others…)

Bonus Tip:

“Belief in external world independent of percipient subject is essential for the success of a metrics program”

Most of Metrics programs that are claimed to be successful stand on this belief.




Anonymous said...

At first, the Bonus Tip was kind of a quick leap, but it may actually the key point.

I do think, however, that you are over-generalizing when you say "Most of Metrics programs that are claimed to be successful stand on this belief"

One of the keys to a successful metrics program - and to nearly every metrics program that *I* have seen as successful - is that all of the stakeholders agree on how the measurements should be perceived, regardless of their interpretation of the world. I guess I'm sorry you've witnessed so many poor metrics programs.

I like your analogies / extensions, but would like to challenge one aspect. Science deals with factual matter, yet many definitions of quality are subjective, and cannot directly be tested against objective data (which is why the metrics analogy makes more sense than the quality analogy). The most difficult thing to do with metrics (and perhaps the hidden point in your post) is to find the set of measurements that accurately "measure quality" for the specified situation - using objective data that has little chance for debate or misinterpretation.

Structure of Scientific Revolutions by Kuhn discusses the phenenomom from the Einstein quote in more depth. The notion has been challenged by some, but mostly on a philosophical level that may or may not be appropriate or accurate (but worth checking out if you want to dive deep on this concept).

Anonymous said...

I just re-read your original post and realized that I had misinterpreted the final quote. Oh well - I guess the comment still has some relevancy.

I'll stop posting comments before coffee.

Gerald M. Weinberg said...

Please note that it is "belief" in the existence of an external world independent of the observer that's essential, not the "existence" of such a world. It might or might not exist, but if we didn't believe it existed, we wouldn't bother to do science (the way we currently do science).

But you don't have to believe in the existence of an "objective" thing called "quality" to do work to improve the satisfaction of people who use a system. You merely have to believe you can measure their "subjective" satisfaction with the system--and not to 7 significant digits, either.

Anonymous said...

Well put - thank you.

The challenge, is indeed to measure "subjective satisfaction", and I think the failure to do this is why a lot of metrics "programs" fail.

Anonymous said...

The quality issue is a mirage. USA big cars never had quality problems. The peaceniks loved Japanese neutrality in the 1960s. When oild prices skyrocketed, US firms sold relabelled small Japanese cars, which sucked, giving the US firms a bad name. The same Japanese firms then sold their better cards direct, bait and switch. And then the declining literacy of USA repairmen made them unable to fix advanced USA cars, telling their customers to buy Japanese. And the leftist cassandras in the universities tried to get revenge on firms who didn't hire them.