39: Captain’s Log
4 March 2013
You know you need to do it, and today Mike tries to convince you. At a minimum errors need to be logged with enough information to point to the line of code, but where do you go from there? Slogging through bug reports, pulling important metrics, and a few bumps and bruises. Plus: The inventory problem developers face, some forgotten glory, defending Yahoo, a batch of your feedback and more!
Hosts
Episode Links
Show Notes:
Feedback
-
Louie points out my super high pitched voice on the word “business” in the last show
-
Mike asks:
“With every platform having its own app store you would think that it would be a boom for Indy Devs. But I don’t believe that is the case because app discovery seems to suck on all platform. Am I crazy or correct? Is there anyone trying to fix this? Like a place that promotes Indy apps?”
-
Dominic’s Question: The Stupid Client Problem
-
Mike share’s the forgotten glory of Balmer doing TV ads in the 80’s
-
A lot of divided opinion re Yahoo’s ban on working from home
Logging
- The essence of ¿Que?
- The feel of ¿Por Que?
Two key types of logging
- Diagnostic logging
Do you care enough to throw an exception up through the app or manage it another way? This is an “it depends” but logging info level messages probably should be skipped.
- Audit logging
Audit logging captures significant events in the system and are what management and the legal eagles are interested in. This is things like who signed off on something, who did what edits, etc. As a sysadmin or developer troubleshooting the system, you’re probably only mildly interested in these. However, in many cases this kind of logging is absolutely part of the transaction and should fail the whole transaction if it can’t be completed.