Rename defn to def and defclass to class

[Kodiologist]

I haven't actually made the changes yet. I opened this PR to check for sufficient agreement before I forge ahead.

Anyway, I'm thinking this is a good time to do this because then defn will get renamed to def and lambda lists will change in the same release, so users can use which keyword is used in a Hy program as a hint as to whether the code has been updated.

[Kodiologist]

To reiterate, the motivation here is better consistency with Python. These special operators are our direct equivalents of the Python keywords, so using the same name helps the user avoid accidentally mixing up Python and Hy conventions when switching between them, as one often needs to do as a Hy programmer. Using a different name from Python makes sense sometimes, for sure, like how we use the shorter fn in place of the clumsy lambda, but there's no clear motivation for defn over def except consistency with different programming languages.

[peaceamongworlds]

I don't really like this idea. As I said in our previous discussion, I just think that using defn and defclass makes a lot more sense, and that, if def exists, it should only be used for the most basic form of assignment, not function definition.

[Kodiologist]

Okay, with you opposed, I would likely need both @allison-casey and @scauligi to agree in order to get this change. What do you guys think? You seemed supportive in #1975.

[Kodiologist]

For what it's worth, we used to have def as a regular assignment form, equivalent to setv, and it caused a lot of predictable confusion, since Python programmers expected it to define a function.

[TheCatster]

I know I'm not a contributor, but in my opinion I don't see a major reason to have Hy even closer to Python. The major reason I was attracted to it was because it was a Lisp I could continue to use even when my work needed Python. Those coming from Clojure and such would have an easier time transitioning, and as someone who just started using the language a week ago and am already fully comfortable, I think I'm an example of that. Python devs wanting to learn Lisp have a smoother transition, as they can still use their Python comforts here while learning the macros and further styles of Lisp. Those are my two cents at least!

[allison-casey]

I've come around to peace's point of view on this one since we last talked about it in #1975. the def prefix are syntactically better constructs than python's def, class, etc forms. It's just less of an issue in python because you can't introduce you're own define forms like you can in a lisp language with macros. We even have it in our own core with defmain, but it's a common idiom all over other lisps and code bases. defmulti in clojure is a macro the creates a new define therefore it gets the def prefix. I do this in my own code all the time too like defcommand when wrapping around discord.py methods. So i don't think this is a good idea anymore. setv makes sense being different to me since it's an assignment capable of mutating the argument. defn and defclass can only introduce/overwrite a define unlike setv

[Kodiologist]

Okay, F to this proposal.