Bart Lateur <[EMAIL PROTECTED]> writes:
>>> C<save>
>>
>>If I had my druthers, save() would be it.
>
>I'm against it. Why? Because it suggests that all it does is save the
>value for later retrieval. It does not: the value is cleared as well. It
>masks the previous global value, as if it didn't exist before.
>
>       save $/ = "\n";
>
>That doesn't quite say it, does it?

What about C<hide> ?

I think C<proxy> or C<deputy> has merit - "while I am out contact ...".

But I still think C<save> is the essence of what it does.

>
>In PostScript, for example, there's a gsave and grestore, the "g"
>standing for graphics state. It allows you to save the current state of
>the graphics output device, on a stack, but it doesn't modify it. 

Neither does local - it is the = "\n" that does the modifying.

>You
>can continue from there. It's a different concept.

But PostScript also has Save/Restore which is a very similar concept
to perl5's local  ;-)

>
>The name "local" is wrong, because it suggests a private value scoped in
>space. 

It is private value scoped in "time"

>It is not, that's lexical values. It is scoped in time. I think
>it's name should refelect that. "temporary"?
>
>       temporary $/ = "\n";
>
>It's a mouthful. But it's going where I'd want it to go.



-- 
Nick Ing-Simmons

Reply via email to