RE:

I would go for:

  • Piece and quite

in the end that's the thing for me. I have tried a lot of tools in the organisational type of range, which are cool but most of it can be substituted with pen, some paper and a watch.

From a architectural perspective

  • more metrics
  • good logging
  • good monitoring

The help to make decisions, even if you only can see correlations in the beginning it helps.

I won't go into tooling or project planning paradigms because they to me are classified as

  • Tools
  • Structure

Those are mainly basics.

Show all replies

Michaela Greiler currently I measure:

  • cyclomatic complexities
  • different loc metrics
  • code coupling
  • server stats
  • coverage
  • call trees
  • bug probabilities

What I would want to measure is throughput / performance so I can correlate code changes with real impacts also the rest of the hardware should get a better historical overview to correlate code changes to real impacts.

Reply to this…

(13 answers) Take me to the question