In Julia, when people use x.f(y), what he actual means is f(x, y) (assuming f relies on x). So there is no need to define f inside a type, just define f(x,y) outside. So, my first conclusion is: it is not a good practice to put function inside a type.
If there is a rule that people can not define function field in a type, then set x.f(y) = f(x,y) does not have the confusion you mentioned early. Also x.f might be good for multi-dispatch. I am not sure how multi-dispatch is implemented, but the prefix x seems provide a good heuristic. Best, Cheng On Thursday, October 8, 2015 at 12:03:45 AM UTC+2, Stefan Karpinski wrote: > > This is a real issue but I don't think that aping the syntax of a very > different paradigm is the right solution. > > On Wednesday, October 7, 2015, <jonatha...@alumni.epfl.ch <javascript:>> > wrote: > >> I think the auto-completion is a bit of an issue, it makes Julia >> functions not easily discoverable. In other language when you have a string >> you can just type dot and tab and see all the methods that you can apply to >> your string, so you almost never need to go to the doc. >> >> Something similar should be possible to implement in Julia with >> methodswith but I'm not sure any IDE support that yet. methodswith might >> also be too slow for that purpose. >> >
