The Do's and Don'ts of Error Handling

Handling errors in programs is tricky.

We want to write systems that work reasonable well even if there are errors in the code. This is possible, and the results have been used in production systems for many years.

This talk outlines the do's and don'ts of error handling. Things you should do if you want to write fault tolerant code, and things you should not do.

I will argue that making fault-tolerant and salable systems is possible if you have got the system architecture right, bit very difficult otherwise.

I'll give examples some of my favorite programming languages of how things should be done, and how they should not be done.