This was great, thanks! I would love to see a comparison of free monads vs tagless final. It seems like they both claim to achieve a similar goal, but it's not clear to me what the advantages of one approach over the other are.
This is so satisfying to watch even though I understand maybe 15% of it, but is this like the Repository Pattern in OOP? Interface the API, and swap with different implementations
so FreeMonads allow you to use natural transformations to change the interpretation, but I still dont get why you cant do that with normal Monads, I guess the fact that the FreeMonad kind of fixes the wrapped type A as a type parameter has to do with it. Thanks for this video, really cool.
I’ll ask one question more. The create/get/create/delete sequence in myLittleProgram has already been executed by the time foldMap is called on myLittleProgram, so a Flatmap’s f must reference lambdas different from those implicit in the myLittleProgram for comprehension. Obviously, this sequence can be executed only once. So what ARE the Flatmap f’s that are called by FoldMap? I’m totally confused!
The CRUD actions are not executed, but merely _described_ by those Free instances. What you get is a "big" data structure, nothing more. When you call foldMap, the natural transformation's functions kick in (those from dbOps2IO).
So free monads, given that "programs" are just monadic descriptions of the execution flow, can be used to independently execute and log stuff. Interesting.
Very elegant code! Some questions: It looks like the execution of myLittleProgram results in a Flatmap(Suspend(ms), f) Free monad for each line, but I don’t see how they are tied together in a single Free monad for the subsequent foldMap to interpret. Further, a Flatmap’s f would encompass all the computations after it in myLittleProgram, so it would seem that a Flatmap’s f would encompass some of the computations of the previous Flatmap f’s, resulting in a repeat of these computations when myLittleProgram is foldMapped. Finally, the get() in myLittleProgram would not produce a result “name”when it’s flatmapped because myLittleProgam has no database to work with when the for comprehension is executed. How does this work? Too many questions? They might not even make sense. A Free monad Part 2 that showed the data structures created during execution would do a lot to explain all that’s going on. Thanks!
hi，Daniel. I want to buy your course on your website, but I find that all of the videos do not have subtitles. I am from a country not using English as mother tongue， so it maybe not that easy for me to understand each word you are saying. Is there any way that I can add subtitles to the courses like on youtube? Thanks.
@@rockthejvm OK thanks for your reply. I'd like to buy a bundle on the Black Friday 2022, will be there any promotion? Or maybe I will buy a monthly membership to have a try first. Thank you anyway.
Just discovered this channel... A real jewel.
Only wrote about 200 lines of Scala in my life, but this channel is brilliant at explaining tough FP concepts. Great work!
Fantastic video, great work explaining free monads!
thank you very much, I've learned a lot of thing, in particular the ~> trick !
This was great, thanks! I would love to see a comparison of free monads vs tagless final. It seems like they both claim to achieve a similar goal, but it's not clear to me what the advantages of one approach over the other are.
Amazing!
This is so satisfying to watch even though I understand maybe 15% of it, but is this like the Repository Pattern in OOP? Interface the API, and swap with different implementations
Most design patterns are different versions of programming to an interface, then swap implementations.
so FreeMonads allow you to use natural transformations to change the interpretation, but I still dont get why you cant do that with normal Monads, I guess the fact that the FreeMonad kind of fixes the wrapped type A as a type parameter has to do with it. Thanks for this video, really cool.
I’ll ask one question more. The create/get/create/delete sequence in myLittleProgram has already been executed by the time foldMap is called on myLittleProgram, so a Flatmap’s f must reference lambdas different from those implicit in the myLittleProgram for comprehension. Obviously, this sequence can be executed only once. So what ARE the Flatmap f’s that are called by FoldMap? I’m totally confused!
The CRUD actions are not executed, but merely _described_ by those Free instances. What you get is a "big" data structure, nothing more. When you call foldMap, the natural transformation's functions kick in (those from dbOps2IO).
Very nice video! Tangential question: what theme are you using? :)
darcula
So free monads, given that "programs" are just monadic descriptions of the execution flow, can be used to independently execute and log stuff. Interesting.
Very elegant code! Some questions: It looks like the execution of myLittleProgram results in a Flatmap(Suspend(ms), f) Free monad for each line, but I don’t see how they are tied together in a single Free monad for the subsequent foldMap to interpret. Further, a Flatmap’s f would encompass all the computations after it in myLittleProgram, so it would seem that a Flatmap’s f would encompass some of the computations of the previous Flatmap f’s, resulting in a repeat of these computations when myLittleProgram is foldMapped. Finally, the get() in myLittleProgram would not produce a result “name”when it’s flatmapped because myLittleProgam has no database to work with when the for comprehension is executed. How does this work?
Too many questions? They might not even make sense. A Free monad Part 2 that showed the data structures created during execution would do a lot to explain all that’s going on. Thanks!
unable to untangle the question :(
kind of advance subject for me, i'll get back later !
In the for comprehension (15:10), Don't you should return "pure( () ) " instead of ()? I mean, I think that you are getting advantage about Unit type
We're using IO.create, and in the block we need to return a value, not another monad.
hi，Daniel. I want to buy your course on your website, but I find that all of the videos do not have subtitles. I am from a country not using English as mother tongue， so it maybe not that easy for me to understand each word you are saying. Is there any way that I can add subtitles to the courses like on youtube? Thanks.
Sadly the Rock the JVM website is not as advanced as TH-cam to add automated captions :(
@@rockthejvm OK thanks for your reply. I'd like to buy a bundle on the Black Friday 2022, will be there any promotion? Or maybe I will buy a monthly membership to have a try first. Thank you anyway.