[lmi] DefectLog: the hammer drops

[Greg Chicares]

In _The Electric Kool-Aid Acid Test_, it was Neal Cassady's wont
to juggle a sledge hammer for hours on end, and Kesey said that
when Cassady (rarely) dropped the hammer, it meant something bad.
I've recently made several mistakes in HEAD, and believe that
means something bad. So it's time to increase self discipline.

After surveying the literature and considering what has worked
well in past years, I developed 'DefectLog', which is now in cvs.
I'd like to ask everyone who writes to HEAD to use it to log any
future regression error. Personally, I'm finding it useful for
certain old errors, too; but I've had more time than anyone else
here to make old errors.

It takes some effort to follow every step. Well, really, what it
takes is resolve. I find it tempting to believe I've transcended
all this, but here's someone who believes he hasn't:

  https://www.tug.org/TUGboat/Articles/tb10-4/tb26knut.pdf

| I recommend that everybody keep an error log such as the one I
| kept for TEX. The amount of extra time required is negligible
| (less than l%), and the resulting records help us to understand
| ourselves and our fallible natures.

Here's his TEX error log:

  http://tex.loria.fr/historique/errorlog.pdf

Here are a few other articles I found interesting:

  
http://www.dcs.qmw.ac.uk/~norman/papers/qa_metrics_article/section_6_problems.html

  http://docserver.fhg.de/iese/2001/reports/072.pdf

  https://www.thedacs.com/techs/defect/defect-toc.php