Episode description

In this episode, we talk about legacy systems: we explore what makes a system "legacy", why engineers often avoid working with them, and what strategies help to effectively work with these systems.

We share our thoughts on how legacy systems have evolved throughout our careers, what is there to learn from them (and operations in general), and why some of these systems are actually very successful systems worth respecting.

We also offer some practical advice to make incremental improvements on legacy systems, using patterns like the "strangler fig" rather than attempting complete rewrites (which are most of the time not the best approach).

Also, don't miss the story of dealing with challenging legacy code (including a jaw-dropping Perl-Java hybrid, and a VB6 mudball of 20000 lines of unreadable business logic code).

Episode Notes:

• Strangler Fig: https://martinfowler.com/bliki/StranglerFigApplication.html/ https://learn.microsoft.com/en-us/azure/architecture/patterns/strangler-fig
• Respect what came before: https://amazon.jobs/content/en/teams/principal-engineering/tenets