Fifteen Years of Successfully Dialyzing Erlang and Elixir Code (Keynote)
summary
A retrospective on fifteen years of using Dialyzer for static analysis of Erlang and Elixir code, covering what Dialyzer checks, how success typing and PLT-based analysis work, practical experiences integrating Dialyzer into developer workflows, and lessons learned about its strengths and limitations.
AI-written summary · about
chapters (17)
- 0:00Start
- 8:56Dynamically typed languages
- 29:37False error reports show-stopper
- 31:01Type system to the rescue?
- 32:51Restroom signs in the U.S.A
- 33:15Tempting conclusion
- 35:48The moral of all this is...
- 38:32Type inference for defect detection
- 39:49The basic question
- 46:29A quick look at inferred domains
- 49:02Function domains revisited
- 49:26Recap
- 50:10Two sides to the story
- 51:39The inference algorithm (PPOP'06)
- 56:14Two more examples
- 58:25Module system to the rescue
- 59:08The length example revisited
