> -----Original Message----- > From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] > On Behalf Of E. Gladyshev > Sent: Wednesday, July 30, 2003 11:57 AM > To: Boost mailing list > Subject: RE: [boost] Re: GUI/GDI template library > > > Every design will have both run-time and > > compile-time aspects. > > I agree. Where do you draw the boundary? What is the > correct mix of these two aspects? That's the tradeoff > I was talking about.
If a decision is based on information that cannot be known at compile-time then you must make that decision at run-time. There is no choice. We don't know the sequence of keystrokes a user will type into an edit control, for instance. If you have enough information to decide something at compile time, do it then - you gain nothing by waiting and gain an increased ability to catch mistakes earlier by doing it immediately. The question then becomes how much run-time flexibility you want to incorporate into your design. With increased flexibility comes increased complexity. For phase (1) of Beman's I would shoot for the bare minimum. Often, we will be able to add in the more flexible run-time logic seamlessly without affecting the compile-time framework: row("name: ", edit(&employee::name); Could later become: row(string_resource(name_string_identifier), edit(&employee::name)); Is this a better answer to your question? Brock _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost