Raymond Hettinger explains Python's dataclasses module, showing the problem it solves and its key design decisions. He places it in a progression from tuples and dicts through named tuples and attrs.
Why Hacker News points people here: HN commenters recommend it as an approachable talk from a respected Python core developer on dataclasses.
“Dataclasses: The code generator to end all code generator”
Hettinger's PyCon talk on dataclasses as a code generator
Full comment
Some of my favorite talks from Strange Loop: * Contracts For Getting More Programs Less Wrong: link * "It's Just Matrix Multiplication": Notation for Weaving: link * Hackett: a metaprogrammable Haskell: link * Git from the Ground Up: link From PyCon: * Beyond Unit Tests: Taking Your Testing to the Next Level: link * Dataclasses: The code generator to end all code generator: this video * Automating Code Quality: link There are probably more I might add later as I remember them.
RecommendskzrdudeMay 2021
“Raymond Hettinger's talks are great.”
Approachable Hettinger talk on dataclasses making Python better at records
Full comment
Raymond Hettinger's talks are great. I know it's not a succint blog post, but they are out there are watching a video is very approachable. Maybe this video (Dataclasses! We could cheekily say Python gets better at something Rust does - dataclasses makes Python better at records.) And link This talk (2013): link Unfortunately this video is now a bit out of touch with modern Python. Another (2013) classic: link it is very well known, and it has the very memorable advice: what's a class that only has one method? That should be a function!
Cites itlunchladydorisMay 2018
“I haven't watched it yet but I saw that Raymond Hettinger presented on data classes at PyCon last week [0].”
Hettinger presented on data classes at PyCon.
Full comment
I haven't watched it yet but I saw that Raymond Hettinger presented on data classes at PyCon last week [0]. [0]: this video
Cites itneulandMay 2018
“he even says that more features from attrs will likely get added, such as data validation [1].”
Hettinger says more attrs features like validation may be added.
Full comment
The standard library is not going to have more advanced features than what the community can do, if only because of the release cycle. Meanwhile, I was not a user of attrs before, rather I used collections.namedtuple and then typing.NamedTuple. So, data classes is a win for me and other people like me who weren't using attrs. And you can continue to use attrs without being bothered at all. In Raymond Hettinger's talk [0] at PyCon 2018, he even says that more features from attrs will likely get added, such as data validation [1]. And that his talk was the one chance for the developer (Eric Smith) to get praise, because everyone is going to instantly second guess every decision he made. [0] this video [1] see 42:36 on future directions
