Re: [gtkmm] libsigc++ is a weird name

["ROUGE Alain" <Alain ROUGE enac fr>]

thanks for your reply.

for a non native english speaker :-)  sigslot was cool. :-(

a slot seems to be a general concept of callback. but in fact it is a 
callback (on function or method).

signal is also strongly connected with posix.

my head hurts.

let's call libsigc++ libsigc++ and sorry for my previous email. 

ar.





Paul Davis <pbd op net>
29/07/02 15:07

 
        Pour :  "ROUGE Alain" <Alain ROUGE enac fr>
        cc :    gtkmm-list gnome org
        Objet : Re: [gtkmm] libsigc++ is a weird name


>libsigc++ deals with signals and slots. 
>sigslot-1.2.tar.gz is a clear name.

"slot" was a term unknown to me until i started using the library. i
still don't like it. as a native english speaker, its ambiguous,
imprecise, and has no strong associated-and-relevant metaphors for
me. i'm not even really sure i like "signal" because of its
connections with POSIX, but i do know that trying to explain what a
"slot" is to a programmer who already understands callbacks is really
quite a task.

in an ideal world, i think we'd be talking, as GTK+ does, about
signals and "handlers"; for most Unix-raised programmers (at least),
this is a pair of english words that have some sort of semantic resonance.

to paraphrase murray (i think):

>libsigc++ implements a typesafe callback system for standard C++. It
>allows you to define signals and then connect handlers to those
>signals. A handler can be any kind of function, either global or a
>member, regardless of whether it is static or virtual. When a signal is
>raised, the handlers are invoked.

--p