Notice the crazier than usual cuts at 08:02. In the old video I said “returns” instead of “takes” but it was at a very very unfortunate moment which made the explanation very confusing 😊 I figured it was better to accept the mistake and just reupload so that I do not create unnecessary confusion 😊
I guess it is much simpler to think first about functions and their input and output parameters. Everything is very intuitive there. When you move to classes and objects you can consider them as set of functions with their parameters and that is why it becomes important to define the direction of the Type parameter just same way as you specify it in the functions.
Christopher I am eternally grateful for the great content you have created, thanks to your videos of design patterns and thoughts of these topics I have grown a lot, hope to see you some time in Seattle! Saludos 🫡 master! 🙇♂️
Here's my invariant: sometimes I'd like to give a bigger thumbs-up than others. This is a good lesson! Worth re-uploading the corrected version, good stuff, can recommend.
I can say that you are one of the best that explain this abstract concept. Thanks for all the passion, you really inspired me and helped in my work. Keep going!!!
I am a Kotlin Student and this made more sense to me than trying understanding it from ChatGPT. Thanks lots mate for the intuitive live example to sink it in!!
Great video Mostachón, I liked how you explained using “general” and “specific” rather than using the terms “wide” and “narrow” as some British devs I follow did
After all the years of watching you, I have to say I am very impressed with your determination to teach Especially when combated with the youtube comments of hatred that everyone deals with!!! Keep up the good work :) You helped me and others I recommended your channel to in my early days so thank you very much
My friends and I watch your videos together whenever you upload a video as such and discuss the ideas. We've never fully understood how generics behave before this video (whether it be in java or in C#). You're great man keep up the new style videos and we're waiting for the LSP sequel video.
Thanks for this video, its really confusing for not native english speaker, but its even more confusing when you try to read about this topic somewhere online! haha. Looking forward for next great videos!
If I take something out of a bag of apples, I expect an apple, not an orange. If it was a bag of fruit, I might get a cherry. You should explain genetic in and out. I love the chopped edited videos.
I really hate that Microsoft made Task instead of ITask. With the latter a a method declared with return type ITask would be able to return a ITask without roundtripping over an extra await.
Thank you so much for this. This is so great. Would love to see you do some type system specific videos (which caters to Functional programming) .. Once again, thanks for the great content
I never saw your type signatures video before. Just watching it now, and yes that's exactly the kind of videos , I would love to see. Although, this might be a niche set of topic. I am also watching the following videos ua-cam.com/play/PLA_-EWSPTJcu4i7RFCl_KeGrrz37C4_Oc.html&si=d02STCKZbOYPCGxM and they are good. But I was wishing you had a playlist of functional programming topics .. something like a crash course of the topics.. ( with some book as a reference, the way you have done for your design pattern videos ) Once again, thanks for covering interesting topics Good to see you uploading new content again.
Hmm I never got to know these principles or forgot them again right after, but I am happy that C++ just prevents me from doing any of this :D I cannot override methods with different signatures and I cannot assign a vector of apples to a vector of fruits :)
Thank you for the question. In my design pattern series I recommend the following two books which are absolute classics: * Design Patterns: Elements of Reusable object oriented software (geni.us/PsXmo) * Head First: Design Patterns (geni.us/nlbA6) I've also got a few videos planned that will make more book recommendations with a bit more context 😊 The design patterns playlist: ua-cam.com/video/v9ejT8FO-7I/v-deo.html
Thanks for explanation but I have a question, if FruitJuicer overrides the method in OrangeJuicer, the interface of method must be same? But the method in fruitjuicer is taking Fruit as input but the parent method is taking Orange type as input?
I honestly never knew this was a thing, I guess I don't think about it, but I also don't think I've ever seen that kind of error before so maybe I just planned things well enough to avoid it. I kinda want to go see if I can break my program doing this now, just because I've never encountered it. I don't know if this is a good practice, but if I have a list of types and I don't know or expect a certain type but can't guarantee it, and I absolutely cannot avoid a situation where I don't know what I'm getting, like a list of Animals that could be Cats, then I'll embed that object with an enum like AnimalType. Then I create the object with the type during construction, and check it to be sure whatever I have is, as your example, a Cat. Anyway, I'd prefer not to have to think so much about it and just be sure I'm not going to do something to a cat that only dogs or ducks are expected to do.
Think of it in terms of substitutability. If a method expects an OrangeJuicer but all you have is a FruitJuicer then you should still be able to pass it since it will always be able to juice oranges. A concrete example is the IComparer interface in C# which is contravariant. If you create a comparer for Fruit you can use that comparer to sort lists of Orange. Even though a list of orange wants an IComparer but all we have is an IComparer. Makes more sense? Thank you for the question.
I was confused with this as well. I hope the following would help. Keep in mind the idea of substitutability from above. From wiki: "A subtype is a datatype that is related to another datatype (the supertype) by some notion of substitutability". It's not always what is 'larger', the context matters. - in the example of covariance we talked about lists, or arrays (consider this as source or output!), it's for them Cat is a subtype of Animal, one can substitute another. An array of cats is a substitution for an array of animals, when speaking in terms of output; - in the example of contravariance we talked about actions, or function parameters (consider this as sink or input!), it's for them FruitJuicer is a subtype of OrangeJuicer, one can substitute another. An action of squeezing out the fruit is a substitution for an action of squeezing out the orange, when speaking in terms of input. P.S. I highly recommend the article on covariance and contravariance in CS on Wikipedia @@gregbell2117
I have yet to wrap my head around bivariance completely. If you have good examples of where it makes sense I would really appreciate it if you would like to share them 😊🙏 😊 Thank you for watching 😊
Thank you so much for making this video. Would you be so kind as to explain why FruitJuicer is a subtype of OrangeJuicer and not the other way around? I mean while the UML diagram follows the statement "FruitJuicer is a subtype of OrangeJuicer" isn't FruitJuicer a superclass of OrangeJuicer in real world?
Hey Chris..thanks for the explanation, I have a question though. How have you defined a "IS A" inheritance relationship between FruitJuicer and OrangeJuice? I dont think that a FruitJuicer "IS A" OrangeJuicer, although a FruitJuicer can juice any fruit(including orange), but how can we say that the FruitJuicer is a sub type of the OrangeJuicer?
great video! I'm wondering, is 'upcasting' and 'downcasting' the way to try to make arrays temporarily covariant/contravariant in runtime (risking a runtime exception)?
Close. Variance regards what is considered a subtype or something else. And if A is a subtype of B then objects of type A can safely be upcast to type B. Covariance and Contravariance regards which type can be considered a subtype of which. So it’s always about upcasting. Thanks for watching 😊
I have a problem statement, I will try to frame it & seems related to the video: My program has 2 things to do - i) Fetch some fruit responses from various places and ii) Transform each fruit responses differently into views Example: I bring AppleResponse and BananaResponse (extends FruitResponse) data by making different API calls (irrelevant detail) I now want to transform AppleResponse and BananaResponse into AppleView1, AppleView2 from AppleResponse & BananaView1 (extends FruitView) from BananaResponse I have an intermediate ResponseStore Map on FruitType { APPLE -> AppleResponse, BANANA -> BananaResponse } ^This is the output of Step-1 : Map
If I understood the video correctly, Java lists could be covariant (since generic types are erased at runtime) but are still are inviariant. Do you know if there's any other reason for this?
Btw, for my programmer brain, what clicked was the explanation from 8:20 to 10:55. Just as feedback if you ever want to flip your explanations to be practical-then-theorical rather than theorical-then-practical. Still awesome stuff :)
Just because you're fed a sequence of Cat to start your declared list of Animals, I don't see why that doesn't allow the more generic list of Animal to accept further things than the Cats it was initialized with 🤨🤔🤔 We aren't trying to put dogs in the Cat list, we're trying to (clearly, as defined) craft an Animal list that we started from a Cat list 🤔 You say "it doesn't support all Animals", while talking about a clearly defined and declared List, so of course it does. Only the List we started our Animal sequence from, does not. 🤨 This is my confusion despite/during your explanation. Thanks!
Thank you for this very important comment! It would be great if you could give examples of words that I am using that are complicated. Either way I will do my best to speak simpler in the future. Thank you again 🙏😊
Didn't expect an analysis of politics here. "Be conservative in what you do." "Be liberal in what you accept from others." Are you telling us that Conservatives tend to be more productive while Liberals tend to be more tolerant of others? Or am I mixing Apples and Oranges?
Haha 😊 Hadn’t thought of it that way. This is a popular saying (stemming from the TCP spec if I’m not mistaken) so I should not take credit for it. Either way we are here using the terms in their literal sense. So we mean conservative as in that we should not return anything wild (meaning that we in the case of TCP should stick to the spec). And we mean liberal as in that we should be open to alternative interpretations (meaning that we in the case of TCP should allow deviations from the spec). Thank you for the comment and for the interesting interpretation 😊. I don’t want to get dragged into politics but my hunch would tell me that I agree with you 😊😊
@@ChristopherOkhravi You're agreeing with yourself which shouldn't be too surprising. Taking this into Game Theory, Contravariance says that it's a Zero-Sum Game ("I win - you lose") or at least "You have to do it my way", while Covariance says it's Win-Win; we can have our Orange and eat it and you can have any Fruit you want.
Reupload because of a mistake in the previous upload. My apologies 🙏
i felt like deja vu :)
May I know what mistake was in the previous upload? I have watched this re-upload and it seems the same as the previous one.
@@NicolasChanCSY I would also like to know specific spots that are different
Notice the crazier than usual cuts at 08:02. In the old video I said “returns” instead of “takes” but it was at a very very unfortunate moment which made the explanation very confusing 😊 I figured it was better to accept the mistake and just reupload so that I do not create unnecessary confusion 😊
It's rare to find a YT channel that explains such content, especially in such an approachable way. Thanks a bunch! 🙏
Thank you 🙏😊
This was an amazing explanation of this. Especially the final part when discussing input and output portion. Great job
Thank you. I'm happy to hear that you appreciate it.
Crazy explanation, bro! No one will forget this concept now.
I guess it is much simpler to think first about functions and their input and output parameters. Everything is very intuitive there.
When you move to classes and objects you can consider them as set of functions with their parameters and that is why it becomes important to define the direction of the Type parameter just same way as you specify it in the functions.
Wow, never heard someone explain that in this way, relating the need of using the out and in directions in classes with functions. Thank you so much!
This concept took me two day to understand when I was learning C#. Thanks for bringing it up, great job.
Literally the only UA-cam channel I use for coding stuff, since college to when I need to brush up on stuff at my job!
Best explanation of in-, co- and contravariance ever. Finally, I get the concept after using it for so many years without fully understandig it.
Christopher I am eternally grateful for the great content you have created, thanks to your videos of design patterns and thoughts of these topics I have grown a lot, hope to see you some time in Seattle!
Saludos 🫡 master! 🙇♂️
Thank you very much 🙏🙏
Here's my invariant: sometimes I'd like to give a bigger thumbs-up than others. This is a good lesson! Worth re-uploading the corrected version, good stuff, can recommend.
Thank you 🙏😊
He is the Hero.. Always explain the concepts with passion and energetic which drives us to understand the concept clearly. Thank you 🎉
Thank you. And thank you for watching 😊🙏
I can say that you are one of the best that explain this abstract concept. Thanks for all the passion, you really inspired me and helped in my work. Keep going!!!
I am a Kotlin Student and this made more sense to me than trying understanding it from ChatGPT. Thanks lots mate for the intuitive live example to sink it in!!
Great video Mostachón, I liked how you explained using “general” and “specific” rather than using the terms “wide” and “narrow” as some British devs I follow did
The best explanation I have even seen including MSDN why List can not be assigned to IList ! Thank you!
Este canal tiene puro oro 🪙
After all the years of watching you, I have to say I am very impressed with your determination to teach Especially when combated with the youtube comments of hatred that everyone deals with!!! Keep up the good work :) You helped me and others I recommended your channel to in my early days so thank you very much
Thank you. I’m very happy to hear that 😊🙏
My friends and I watch your videos together whenever you upload a video as such and discuss the ideas.
We've never fully understood how generics behave before this video (whether it be in java or in C#).
You're great man keep up the new style videos and we're waiting for the LSP sequel video.
Wow. That’s awesome. Makes me very happy to hear that 😊
Thanks for this video, its really confusing for not native english speaker, but its even more confusing when you try to read about this topic somewhere online! haha. Looking forward for next great videos!
Simply Amazing Explanation!
If I take something out of a bag of apples, I expect an apple, not an orange. If it was a bag of fruit, I might get a cherry. You should explain genetic in and out.
I love the chopped edited videos.
Absolutely amazing explanation. Your content is always on top! Always easy to recomment to anyone. Thanks for the effort🎉❤
Amazing presentation, so clear
Wow, I loved your examples at the beginning!
Thanks for sharing this valuable lesson 🎉
I really hate that Microsoft made Task instead of ITask. With the latter a a method declared with return type ITask would be able to return a ITask without roundtripping over an extra await.
great explanation! Thanks bro.
Very well explanation. It was amazing, thank you!!
Amazing work! Great examples and explanation!
Your content is amazing. Great stuff
great explanation!
Thank you so much for this. This is so great. Would love to see you do some type system specific videos (which caters to Functional programming) .. Once again, thanks for the great content
Thank you very much for the feedback. 😊🙏 Are you the thinking along the lines of my series on Ramda JS or more like my video on type signatures?
I never saw your type signatures video before. Just watching it now, and yes that's exactly the kind of videos , I would love to see. Although, this might be a niche set of topic.
I am also watching the following videos
ua-cam.com/play/PLA_-EWSPTJcu4i7RFCl_KeGrrz37C4_Oc.html&si=d02STCKZbOYPCGxM
and they are good. But I was wishing you had a playlist of functional programming topics .. something like a crash course of the topics..
( with some book as a reference, the way you have done for your design pattern videos )
Once again, thanks for covering interesting topics
Good to see you uploading new content again.
Hmm I never got to know these principles or forgot them again right after, but I am happy that C++ just prevents me from doing any of this :D I cannot override methods with different signatures and I cannot assign a vector of apples to a vector of fruits :)
Amazing content.
You're amazing, keep up the good work!
Thank you for video! Could you share some books about OOP, design patterns, SOLID, etc, that helped you to understand these themes.
Thank you for the question. In my design pattern series I recommend the following two books which are absolute classics:
* Design Patterns: Elements of Reusable object oriented software (geni.us/PsXmo)
* Head First: Design Patterns (geni.us/nlbA6)
I've also got a few videos planned that will make more book recommendations with a bit more context 😊
The design patterns playlist:
ua-cam.com/video/v9ejT8FO-7I/v-deo.html
Thanks@@ChristopherOkhravi
@@ChristopherOkhraviThe geni links are not valid due to ) in the formatting
Very enlightening video, I think I've a better understanding now about Producer Extends Consumer Super (PECS) rule applied in Java 😀
Sorry, you lost me at “Fruit Juicer” is a subtype of “Orange Juicer” ! Nope…. Not in my world 😮
Thanks for explanation but I have a question, if FruitJuicer overrides the method in OrangeJuicer, the interface of method must be same? But the method in fruitjuicer is taking Fruit as input but the parent method is taking Orange type as input?
I thought the whole contravariance was that the method signatures don't have to be the same, based on the type hierarchy?
Back to the drawing board... Gotta figure this out with pen and paper.
Great video!
I honestly never knew this was a thing, I guess I don't think about it, but I also don't think I've ever seen that kind of error before so maybe I just planned things well enough to avoid it. I kinda want to go see if I can break my program doing this now, just because I've never encountered it. I don't know if this is a good practice, but if I have a list of types and I don't know or expect a certain type but can't guarantee it, and I absolutely cannot avoid a situation where I don't know what I'm getting, like a list of Animals that could be Cats, then I'll embed that object with an enum like AnimalType. Then I create the object with the type during construction, and check it to be sure whatever I have is, as your example, a Cat. Anyway, I'd prefer not to have to think so much about it and just be sure I'm not going to do something to a cat that only dogs or ducks are expected to do.
you are just amazing 😇
Great one!!!
how come FruitJuicer is a subtype of OrangeJuicer?? 4:37
Think of it in terms of substitutability. If a method expects an OrangeJuicer but all you have is a FruitJuicer then you should still be able to pass it since it will always be able to juice oranges.
A concrete example is the IComparer interface in C# which is contravariant. If you create a comparer for Fruit you can use that comparer to sort lists of Orange. Even though a list of orange wants an IComparer but all we have is an IComparer.
Makes more sense?
Thank you for the question.
@@ChristopherOkhraviI understand your explanation but am still confused - aren't subtypes usually more specific?
@@ChristopherOkhravi thank you bud you are the best !
I was confused with this as well. I hope the following would help. Keep in mind the idea of substitutability from above.
From wiki: "A subtype is a datatype that is related to another datatype (the supertype) by some notion of substitutability". It's not always what is 'larger', the context matters.
- in the example of covariance we talked about lists, or arrays (consider this as source or output!), it's for them Cat is a subtype of Animal, one can substitute another. An array of cats is a substitution for an array of animals, when speaking in terms of output;
- in the example of contravariance we talked about actions, or function parameters (consider this as sink or input!), it's for them FruitJuicer is a subtype of OrangeJuicer, one can substitute another. An action of squeezing out the fruit is a substitution for an action of squeezing out the orange, when speaking in terms of input.
P.S. I highly recommend the article on covariance and contravariance in CS on Wikipedia @@gregbell2117
Yes. I am struggling with that inversion too.
Are there any plans to start a series of videos about Domain Driven Design?
I didn't even know there were words for these concepts.
Nice.
Amazing
Bivariance makes sense in a few cases, bith theoretically and also practically.
I have yet to wrap my head around bivariance completely. If you have good examples of where it makes sense I would really appreciate it if you would like to share them 😊🙏 😊 Thank you for watching 😊
Thank you so much for making this video. Would you be so kind as to explain why FruitJuicer is a subtype of OrangeJuicer and not the other way around? I mean while the UML diagram follows the statement "FruitJuicer is a subtype of OrangeJuicer" isn't FruitJuicer a superclass of OrangeJuicer in real world?
Hey Chris..thanks for the explanation, I have a question though. How have you defined a "IS A" inheritance relationship between FruitJuicer and OrangeJuice? I dont think that a FruitJuicer "IS A" OrangeJuicer, although a FruitJuicer can juice any fruit(including orange), but how can we say that the FruitJuicer is a sub type of the OrangeJuicer?
💚
great video! I'm wondering, is 'upcasting' and 'downcasting' the way to try to make arrays temporarily covariant/contravariant in runtime (risking a runtime exception)?
Close. Variance regards what is considered a subtype or something else. And if A is a subtype of B then objects of type A can safely be upcast to type B. Covariance and Contravariance regards which type can be considered a subtype of which. So it’s always about upcasting. Thanks for watching 😊
Earned a sub
Welcome 😊 and thank you 🙏
I have a problem statement, I will try to frame it & seems related to the video:
My program has 2 things to do -
i) Fetch some fruit responses from various places and
ii) Transform each fruit responses differently into views
Example: I bring AppleResponse and BananaResponse (extends FruitResponse) data by making different API calls (irrelevant detail)
I now want to transform AppleResponse and BananaResponse into
AppleView1, AppleView2 from AppleResponse & BananaView1 (extends FruitView) from BananaResponse
I have an intermediate ResponseStore Map on FruitType { APPLE -> AppleResponse, BANANA -> BananaResponse }
^This is the output of Step-1 : Map
Can you do more functional programming
Gladly! 😊😊
I know this is undoubtedly the best explanation but still am confused 😂
If I understood the video correctly, Java lists could be covariant (since generic types are erased at runtime) but are still are inviariant. Do you know if there's any other reason for this?
Btw, for my programmer brain, what clicked was the explanation from 8:20 to 10:55.
Just as feedback if you ever want to flip your explanations to be practical-then-theorical rather than theorical-then-practical. Still awesome stuff :)
Very valuable feedback. Thank you very much 🙏😊
How many attempts at the orange juggle before camera perfect? 😄
Don’t underestimate my juggling skills 😁😁😁
Why did you reupload this?
There was a mistake in the previous upload that makes it too confusing. My apologies! 🙏
@@ChristopherOkhravi
Seems the same video to me. I watched both of them 😅
so contravariance === generics?
The ending 😂
5:35
Just because you're fed a sequence of Cat to start your declared list of Animals, I don't see why that doesn't allow the more generic list of Animal to accept further things than the Cats it was initialized with 🤨🤔🤔 We aren't trying to put dogs in the Cat list, we're trying to (clearly, as defined) craft an Animal list that we started from a Cat list 🤔 You say "it doesn't support all Animals", while talking about a clearly defined and declared List, so of course it does. Only the List we started our Animal sequence from, does not. 🤨 This is my confusion despite/during your explanation. Thanks!
Ye that is too confusing
Me *doing leonardo de caprio pointing at tv* : pssst psssst pssst yup that is me
Bag of fruit has apples, bananas, grapes, oranges.....
I am not a native english speaker, can you replace some words with simpler variants
Thank you for this very important comment! It would be great if you could give examples of words that I am using that are complicated. Either way I will do my best to speak simpler in the future. Thank you again 🙏😊
@@ChristopherOkhraviYour video was perfect. I wonder if this comment was a joke 😂
@@jay31415oh 😬😊🤦♂️
You just have to practice some coding to be good in the language. Then come on this channel to ace in every project or work place.
Didn't expect an analysis of politics here. "Be conservative in what you do." "Be liberal in what you accept from others." Are you telling us that Conservatives tend to be more productive while Liberals tend to be more tolerant of others? Or am I mixing Apples and Oranges?
Haha 😊 Hadn’t thought of it that way. This is a popular saying (stemming from the TCP spec if I’m not mistaken) so I should not take credit for it. Either way we are here using the terms in their literal sense. So we mean conservative as in that we should not return anything wild (meaning that we in the case of TCP should stick to the spec). And we mean liberal as in that we should be open to alternative interpretations (meaning that we in the case of TCP should allow deviations from the spec).
Thank you for the comment and for the interesting interpretation 😊. I don’t want to get dragged into politics but my hunch would tell me that I agree with you 😊😊
@@ChristopherOkhravi You're agreeing with yourself which shouldn't be too surprising.
Taking this into Game Theory, Contravariance says that it's a Zero-Sum Game ("I win - you lose") or at least "You have to do it my way", while Covariance says it's Win-Win; we can have our Orange and eat it and you can have any Fruit you want.