Is there a doc for naming convention?

[sth4nth]

Hi, is there a document for naming convention for Julia function names?

[Viral Shah]

There are informal conventions that we follow, where type names are in camel case, and function names are all lower case, with underscores. 

-viral

[Stefan Karpinski]

Jeff hates underscores and as a result most things in Base don't use them. But for long variable names, not using underscores gets pretty out of hand. I was considering having a convention that names in Base, as "primitives" in some sense, have no underscores whereas library code would use the underscore convention. Not sure if that's going to fly, but I have a hard time swallowing just smashing all words together everywhere.

--
 
 
 

[John Cowan]

On Thu, Oct 11, 2012 at 11:36 AM, Stefan Karpinski <ste...@karpinski.org> wrote:
> Jeff hates underscores and as a result most things in Base don't use them.

I'm not a fan either. For one thing, when you have underlined
hyperlinks in documentation, the underscores in names disappear.

> But for long variable names, not using underscores gets pretty out of hand.
> I was considering having a convention that names in Base, as "primitives" in
> some sense, have no underscores whereas library code would use the
> underscore convention.

I'd rather use the lowerCamelCase convention for long method names,
since we are already using UpperCamelCase for types.

> Not sure if that's going to fly, but I have a hard
> time swallowing just smashing all words together everywhere.

Yes, thatisnotgoodeither.

--
GMail doesn't have rotating .sigs, but you can see mine at
http://www.ccil.org/~cowan/signatures

[BobC]

Well, Julia could use a non-alpha 'word-name-joiner' character other than '_' to avoid the underlining problem.

IIRC, didn't DEC use '$' for that purpose (in VMS filenames)?  Julia could too!  my$Julia$variable$name = value

It's been too many decades since I used UPPERCASE-ONLY FORTRAN, but I have a whiff of a recollection that we used a joiner character there too...

-BobC

[Douglas Bates]

On Thursday, October 11, 2012 11:32:42 AM UTC-5, BobC wrote:

Well, Julia could use a non-alpha 'word-name-joiner' character other than '_' to avoid the underlining problem.

IIRC, didn't DEC use '$' for that purpose (in VMS filenames)?  Julia could too!  my$Julia$variable$name = value

It's been too many decades since I used UPPERCASE-ONLY FORTRAN, but I have a whiff of a recollection that we used a joiner character there too...

I think we may be wandering a bit off-topic here but wasn't the ancient FORTRAN convention that you could insert blanks in an identifier?  Of course, there was a maximum of 6 non-blank characters in an identifier so you couldn't exactly write "War and Peace" while constructing an identifier. 

[BobC]

Duh.  Julia is Unicode: There are a gazillion useful 'joiner' characters.  Pick one (perhaps '˽'), and assign it to Meta/Alt-'_' on your keyboard.

-BobC

[Patrick O'Leary]

You're assuming there's always a sufficiently-featured editor that you control available, and that your whole workflow is UTF-8 safe. Even if that's true, setting up a keyboard shortcut seems like a pretty silly hurdle to just being able to use Julia effectively (assuming we're talking about things in Base or publicly-shared packages , here.)

(Personally, the fact that you can use dashes in identifiers is one of the things I appreciate about Lisps.)

[Tim Holy]

On Thursday, October 11, 2012 10:04:43 AM Patrick O'Leary wrote:
> (Personally, the fact that you can use dashes in identifiers is one of the
> things I appreciate about Lisps.)

I vote for *. That's what we use for string concatenation, and so it means
concatenate the name.

Oh wait, you don't think that will risk any confusion with multiplication, do
you? Drat...

--Tim

[Patrick O'Leary]

On Thursday, October 11, 2012 12:12:34 PM UTC-5, Tim wrote:

On Thursday, October 11, 2012 10:04:43 AM Patrick O'Leary wrote:
> (Personally, the fact that you can use dashes in identifiers is one of the
> things I appreciate about Lisps.)

I vote for *. That's what we use for string concatenation, and so it means
concatenate the name.

Brilliant.
 

Oh wait, you don't think that will risk any confusion with multiplication, do
you? Drat...

(*)(a::Symbol, b::Symbol) = symbol(string(a)*string(b))

eval(:this*:is*:a*:long*:variable*:name)

:D

[Harlan Harris]

Unicode's pretty cumbersome to type, and risky to read. But there are plenty of unused ASCII characters! I vote for Backspace. The words in your identifiers will literally overlap.

 -Harlan

--
 
 
 

[Patrick O'Leary]

On Thursday, October 11, 2012 12:19:27 PM UTC-5, Patrick O'Leary wrote:

eval(:this*:is*:a*:long*:variable*:name)

:D

(*)(a::Symbol, b::Symbol) = symbol(string(a)*string(b))

 More from the "don't try this at home" dept.

julia> macro v_str(s)
       symbol(s)
       end

julia> v"this is an arbitrary name"
this is an arbitrary name not defined

julia> v"this is an arbitrary name" = 2
2

julia> v"this is an arbitrary name"
2

[Jasper]

Why dont we use the Julia set as separator. We just need to ask the
Unicode consortium to add it. It would be great publicity!

[Jeffrey Sarnoff]

there is an open issue about adopting the convention: postfix '!' to mutable function's names 
https://github.com/JuliaLang/julia/issues/907 (discussion at https://groups.google.com/forum/#!topic/julia-dev/W3vqpy0fSMs/discussion)

I am for it.  And I am using the postfix "!!" for dangerous (unchecked, &tc) functions.