Re: Accessor methods and (auto)release: conclusion
Re: Accessor methods and (auto)release: conclusion
- Subject: Re: Accessor methods and (auto)release: conclusion
- From: Kirk Kerekes <email@hidden>
- Date: Tue, 6 Aug 2002 12:26:34 -0500
Just a thought --
Perhaps both the "handle all the possibilities, no matter how arcane" and
"keep it simple" camps have a point.
In the software development cycle there are really two entirely different
audiences: the developer/tester, and the customer.
The developer/tester needs the software to "break" where it is breakable,
quickly, definitively and in a manner that lends itself to being easily
figured out.
The customer doesn't want the software to break at all.
So it seems to me that there would be a point in accessors that were
sensitive to the build style -- a development build should use the most
delicate (and probably simplest) accessor pattern -- possibly even a
somewhat wrong-headed one, just to reveal issues of thread safety, etc. The
goal of testing is, after all, to break the software before the customer
ever sees it, in the hopes of making it more break-resistant.
The release build should, on the other hand, probably use a more robust
accessor pattern that may forestall crashes caused by unanticipated user
action ("... because fools are so ingenious").
So I would suggest that perhaps accessors with a built-in
conditional-compile-switch might be a useful pattern.
_______________________________________________
cocoa-dev mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/cocoa-dev
Do not post admin requests to the list. They will be ignored.