Type Hints - Guido van Rossum - PyCon 2015

[@jk3089]

OMG! I have never been expecting I'll watch Guido Van Rossam

[@WasiMaster]

At 28:04. The better world example now works as of python 3.9 :)

[@MaikSpike]

Guido is the best thing ever to happen to Python

[@Krasbin]

+Miguel ‘Maikflow’ Pena And without Guido there would be no Python. He sorta kinda started it all.

[@filiphron3147]

+Miguel „Maikflow“ Pena Python is the best think that happened to Guido.

[@msshroff]

actually pyhon is the best thing ever to happen to guido :)

[@sadhlife]

here's the right quote: Guido is Python

[@Twisted_Code]

thank you for this awesome feature I didn't realize existed until about two weeks ago. I've already preemptively caught at least one bug thanks to my IDE using a type hint I included. I imagine it won't be the last.

[@megalowhale8175]

"Some bugs get found sooner if you have type annotations." Understatement of the century. It completely eliminates an entire class of errors. Personally, when languages come out that DON'T feature type-checking, I really feel like this is regressive and taking a step in the wrong direction.

[@Konstruktivismus]

I am not so sure about that. If you write tests for all parts of your code (coverage=100%), then you also have some sort of type checking of the whole code. And actually this testing removes many other types of errors as well, and therefore is just necessary to write stable and good code.

[@megalowhale8175]

Konstruktivismus​ unit testing alone is NOT enough. There are many possible type errors which remain possible even with 100% coverage. This becomes more apparent the less trivial the code base is. Look up the article by Evan Farrer on unit testing + static typing, as he provides a number of examples of this.

[@garlic-os]

Can I get a picture of you with that mango thing?

[@Twisted_Code]

12:15 "legacy code you don't want to change" yeah... as the saying goes, if it isn't broke, don't *_touch_* it!

[@nikolayyotsov1240]

what happen to "Explicit is better than implicit. " ?

[@robbert-janmerk6783]

I would rather ask: what happened to "There should be one-- and preferably only one --obvious way to do it."

[@sadhlife]

this is explicit :0

[@sadhlife]

@@robbert-janmerk6783 I'll be honest. writing typed code is the only obvious way to write most Python code now

[@legion_prex3650]

@@sadhlife WORD!

[@astroboy512]

@@robbert-janmerk6783 "preferably only one", note the "preferebly"

[@nikosc]

is there intention from any popular open source python libraries (e.g. django, pyramid, sqlalchemy..) to use type hints in their codebase?

[@ovangle]

Nikos C. The simple answer would be no, not in the short term, because they're all stuck using six libraries to support users on py2.

[@AistisJokubauskas]

+Thomas Stephenson well I don't quiet understand your statement. If I understood correctly type information goes into pyi file, which is not interpreted by python 2. Would it be correct?

[@VaibhavMishraVM]

+Kiki Klank yes according to the slides.

[@tomaszmakuch]

Guy who edited the video had one job: NOT to cover the slides with little Guido...

[@calebparks8318]

No offense, but C++ saw this coming from before.

[@ashrasmun1]

imo everyone who disregards the power of types is just irresponsible