David Deutsch wrote:
> Good question! (Also: shame on me for not finding those guidelines ealier
> O_o .../awkward/)
> 
> I think the basic recommendation that PSR-1 makes on this are pretty sound:
> Make a choice and stick with it. As a personal preference, I use StudlyCaps
> for classes (this is a 'must' in PSR-1), camelCase for methods and
> under_scores for variables. I find that that helps separate the three
> concepts, even though it might seem like "hey, why do you use three ways of
> doing something". Because it's three separate concepts!

We made our choice and we're going to stick with it. Thus use underlines
and no caps. And we're not doing PSR-1 here.
> 
> From what I've seen in the plugins so far, it seems like RCM pretty
> consistently uses lowercase_underscored for all three.
> 
> Sidenote: My IDE actually gives me a lot of funny warnings about the class
> names in RCM - something along the lines of "there is probably a conceptual
> error here, somewhere" I always imagine it standing before me with eyes
> twitching "dude, this aint right!" ;-)

I'm wondering what your IDE means by "conceptual errors".
> 
> Sidenote 2: The Guidelines references PEAR standards. I used to really
> dislike those, particularly for the ~80 character line limit, which seemed
> silly ("Oooh, your b/w terminal can only do 80 characters per line!?").
> Fast forward a couple of years and I absolutely love an 80 char line limit.
> It's like an addiction to shortness.* There are a couple of outdated things
> in there and I might forever think that 4 spaces is silly, but yeah, not as
> bad today ;-)

Agreed. After this cleanup we should definitely rewrite the guidelines.
> 
> *Sidenote 3: This also helped me come around on Namespaces, really. "use...
> as" REALLY helps keep lines short when using classes and particularly when
> calling static methods.

We're not there yet.

~Thomas
_______________________________________________
Roundcube Development discussion mailing list
[email protected]
http://lists.roundcube.net/mailman/listinfo/dev

Reply via email to