[cgiapp] ($self vs $c) was: Re: New Plugin: RunmodeDeclare

[George Hartzell]

Mark Stosberg writes:
 > On Mon, 29 Sep 2008 00:18:02 +0100
 > Richard Jones <ra.jones at dpw.clara.co.uk> wrote:
 > 
 > > Mark Stosberg wrote:
 > > > As Rhesa proposed, he has followed through and released a plugin to add the
 > > > Devel::Declare / Method::Signatures magic to CGI::Application:
 > > > 
 > > > http://search.cpan.org/perldoc?CGI::Application::Plugin::RunmodeDeclare
 > > 
 > > Looks like an interesting alternative to CAP::AutoRunmode. One minor 
 > > irritant is having to put 'runmode foo ($c:)' everywhere (only if using 
 > > $c instead of $self of course). Might be useful to be able to declare 
 > > use CGI::Application::Plugin::RunmodeDeclare(invocant => '$c'), or just 
 > > use CAP::RunmodeDeclare qw($c).
 > 
 > I agree with this sentiment.
 > 
 > This brings back to life the discussion we were having on IRC the other
 > day about "$self" vs "$c".  
 > 
 > I like "$c" because it's shorter and since it's used all over the code
 > base, it's clear enough what it is. And, the first element to 'shift'
 > off of @_ in a method must be the invocant. 
 > 
 > Michael Peters and Richardo argued in favor of "$self" citing the
 > "Principle of Least Surprise" and the expectation that this non-standard
 > change would generate more harm in confusion than benefits in
 > keystrokes. 
 > 
 > Although I still prefer "$c" in my own code, I thought they made good
 > points, so I switched back from "$c" to "$self" in the docs and code of
 > the latest CGI::App dev release. 
 > 
 > But I'm interested in more opinions on the issue. Richard, do you
 > particularlly prefer "$c", or were you mostly following along the docs
 > that were updated to switch from "$self" to "$c" ?

I'll toss in one vote for $self being much less Surprising, both to me
and to folks with whom I'm sharing code.

g.