I've been saying for a few years now that "Computer Science" is a total misnomer... it's far more akin to "Computer Philosophy" or in many cases "Computer Fanatical Ideology".... but watching Kevlin pull apart unit tests what we can see is "The Scientific Method" for the first time what we're seeing is some real Computer SCIENCE.
Computer Semiotics is interesting to cover the concepts being employed All software is hallucinatory, subjective and easily open to interpretation The art is reducing the complexity to comprehensibility by humans
If you can remember that code is used to convey intention, not just make the computer work, then you have the concept Write code as if you can never be asked about it and answer the question “why is this hete?” And “how does this work?”
Classic kevlin talk. Way too long for the content being presented, but pretty good one you get to the meaty stuff. Mentioning "Word friday", his facebook page. Photos of books. Lots of quotes from various sources, including from the book he edited and slides not in sync with what's he's saying (probably not his fault).
These conferences seem afraid to just have a talk that's like 15 minutes where you're presented the idea, have a chance to understand it, and then just go on with your day.
The single assert per test rule is ABSURD. Post-condition could consist of multiple qualities, each is represented by a `assert` statement. And yes, they do blindly apply to this rule.
This is so bad on so many levels... when he talks about test for multiple classes, I mean it's a unit test it should test a unit hence a single class, it's not b u t it's g u t. What you mean is you need a different level of testing to test the interactions between classes and these tests ate not unit tests
If I have a clas Foo that calls functions on another class Bar and I want to unit-test Foo. How do I do it whitout making my code horrible or going towards an anti-pattern? You want to test a functionality, not a class, function or interaction between one or two classes. Also, just like he said. If you revert to a rule as a truth you're doing it wrong. A rule has a reason to exist in a specific contest of the domain of the problem it is supposed to solve. Outside of it, it's useless.
blipman17 I could argue about what shade of blue the sky is, but I actually contacted kevlin on social media to ask him if he had an opinion and he referred me to an article he wrote on the subject and I now agree with what he said. Here is the link: www.theregister.co.uk/2007/07/28/what_are_your_units/
blipman17 I don't think I am colour blind, so I'd say yup sky is blue (its kinda gray here actually cause it's raining but you know its blue most of the time XD)
I've been saying for a few years now that "Computer Science" is a total misnomer... it's far more akin to "Computer Philosophy" or in many cases "Computer Fanatical Ideology".... but watching Kevlin pull apart unit tests what we can see is "The Scientific Method" for the first time what we're seeing is some real Computer SCIENCE.
Computer Semiotics is interesting to cover the concepts being employed
All software is hallucinatory, subjective and easily open to interpretation
The art is reducing the complexity to comprehensibility by humans
If I tried to keep all of this philosophy in my head as I work my head would explode.
If you can remember that code is used to convey intention, not just make the computer work, then you have the concept
Write code as if you can never be asked about it and answer the question “why is this hete?” And “how does this work?”
Why do the slides get all wonkey in so many of these videos?
Here are the slides for people to view while watching: www.slideshare.net/Kevlin/programming-with-guts-41938796 .
"... Oh yes, functions." I'm officially a fan.
23:20 is where the set up for that joke starts. Great stuff. Henney is subtly hilarious in the most arrogantly academic ways. I love it.
Classic kevlin talk. Way too long for the content being presented, but pretty good one you get to the meaty stuff. Mentioning "Word friday", his facebook page. Photos of books. Lots of quotes from various sources, including from the book he edited and slides not in sync with what's he's saying (probably not his fault).
He's frustating speaker like that, he has interesting ideas covered in all kinds of useless nonsense.
These conferences seem afraid to just have a talk that's like 15 minutes where you're presented the idea, have a chance to understand it, and then just go on with your day.
Bob "Uncle" Martin telling people what they are and aren't allowed to do hasn't aged well
The single assert per test rule is ABSURD. Post-condition could consist of multiple qualities, each is represented by a `assert` statement. And yes, they do blindly apply to this rule.
This is so bad on so many levels... when he talks about test for multiple classes, I mean it's a unit test it should test a unit hence a single class, it's not b u t it's g u t. What you mean is you need a different level of testing to test the interactions between classes and these tests ate not unit tests
If I have a clas Foo that calls functions on another class Bar and I want to unit-test Foo. How do I do it whitout making my code horrible or going towards an anti-pattern? You want to test a functionality, not a class, function or interaction between one or two classes. Also, just like he said. If you revert to a rule as a truth you're doing it wrong. A rule has a reason to exist in a specific contest of the domain of the problem it is supposed to solve. Outside of it, it's useless.
blipman17 I could argue about what shade of blue the sky is, but I actually contacted kevlin on social media to ask him if he had an opinion and he referred me to an article he wrote on the subject and I now agree with what he said. Here is the link: www.theregister.co.uk/2007/07/28/what_are_your_units/
Sébastien Lalonde That makes sense. :) ... The sky is blue, right?
blipman17 I don't think I am colour blind, so I'd say yup sky is blue (its kinda gray here actually cause it's raining but you know its blue most of the time XD)
Sébastien Lalonde a Unit can be package or artifact as well