Bugfix of Damocles

There are a few things that crop up in software development, and I am sure in other engineering.

I am sure some people have heard of is the Heisenbug. Which is a play on words related to Heisenberg uncertainty principle, which, in simplest terms, means that when you try to measure/observe something, you change it. In software a Heisenbug is a bug which goes away (or moves) when you add code to try and debug or measure it. They are not that common, but when they happen they are really a pain in the neck.

So today I want to coin a new phrase (and I bet I am not the first), the Bugfix of Damocles. Like the Sword of Damocles, with the risk of something happening any moment.

This particular bug fix is a hopeful fix to a bug, one you really think will fix the problem but cannot, for various reasons, be sure. This happens notably when you have an intermittent problem, perhaps one that takes days or weeks to manifest, and even worse when it is a problem that only happens on a live system. So you have no choice but to try the bug fix on real customers.

  • It will take days, weeks, or months to know if the bug fix works and fixes the bug.
  • However, you may find that it does not fix it at any moment (i.e. when the bug happens, the system crashes, or whatever).

It is a stressful time.

I won't go in to details on the latest Bugfix of Damocles that is stressing me, and my team, as they will be doing an awesome blog post on this, and the previous issue, both of which are really obscure hardware issues that we are working around (we hope).

But, suffice to say, I have made a new mode for my e-paper signs (thanks for suggestion Daniel).

P.S. The sign now uses SNMP to update automatically :-)

No comments:

Post a Comment

Comments are moderated purely to filter out obvious spam, but it means they may not show immediately.

TOTSCO changing the rules again

One of the big issues I had in initial coding was the use of correlationID on messages. The test cases showed it being used the same on a se...