[10x24] What does "Multiple Dispatch" mean in Julia?

[@amj864]

Bro, this was the simplest clearest definition of multiple dispatch.

[@ughnicewatch]

I keep coming back to your videos again and again. You're my default reference for Julia. Thanks for making these.

[@doggodotjl]

You're welcome. Thanks for watching!

[@GuanglinDu]

Concise and to the very point! Well done.

[@doggodotjl]

Thanks!

[@mehdimabed4125]

Cool video (and series in general ^^), very clearly explained, but I don't get th ebenfit of using multiple dispatch over function overloading... In your example, why using "mygenericfunction(1)" instead of "mytypeof(1)" directly ? Thanks :)

[@doggodotjl]

My example probably wasn't the best. Rather than try to answer you're question, here's a link to Stefan Karpinski's JuliaCon 2018 presentation on Multiple Dispatch. He does a much better job of explaining the benefits of Multiple Dispatch: kzbin.info/www/bejne/oZSceaqprapkhbs

[@tonl4498]

Thanks for the video, it helps me a lot in learning Julia. Can you please do a video on function parametric type too?

[@johnbolt2686]

Thank you very nice video.

[@cemtutum273]

@doggodotjl Thanks a lot for the videos! Multiple dispatch was a bit confusing for me. It sounds like a function overloading to me, but it will get more clear after some time while looking into more examples. Is this replacing the following scenario: For instance, I can write a function with an (generic) argument, x, without a data type definition and have multiple if-loops checking if typeof(x) == Integer, or typeof(x) == String, etc. Instead of adding new "elseif"s for every new case, each user add a function with the same name (i.e., another method basically) without touching the original function (or methods) from the base library?... It almost sounds like "inheritance" in Python when we use "super()", correct?? SO, does Julia replicate OOP capabilities with multiple dispatch???

[@Psy8cho]

I would like to know this as well :)

[@Lemmy4555]

i like the concept of multiple dispatch, but sometimes can be a pain. For instance, how would you create a custom custom display function that always display the full DataFrame? Intially i though of just overloading the "show", the idea was to call the original "show" function with a predefined set of parameter, but i found out this is very difficult to do. The second approach i've tried was to create a Struct FullDisplay that took the DataFrame in input and i've create a "show" method for that struct, but i didn't like how it looked. The third approach was to just create a new "custom_display" function that calls show with the parameters i want, but i'm not happy with having both display and custom_display. In OOP I would have solved this easily by extending the class and then calling an hypothetical super.print in the custom defined print function.

[@TJ-hs1qm]

javascript is OO? You meant to say Java or Smalltalk 😀