Re: Renaissance tutorial

[Nicola Pero]

> > [...NSTopLevelObjects for getting the top level objects from a
> > gsmarkup/nib...]
> >
> >It seems you are suggesting a different way of letting the program get the
> >list of the top-level objects - that if the file owner responds to a
> >certain method, say
> >
> > - (void) didLoadObjectsFromGSMarkup: (NSArray *)topLevelObjects;
> >
> >then it's called, with the top level objects as argument in an array.
> >
> >I quite like your proposal :-) it frankly seems better than the nib hack.
> >
> >Any suggestion for a good method name ?
> >
> Well I'm a bit torn, on the one hand I dislike offering multiple ways to 
> achieve the same goal,

It's not necessarily a bad thing, if done discretely.


> on the other hand this is undocumented and therefore is (theoretically)
> more vulnerable to change

It's undocumented for nibs, but it's documented in Renaissance.  The
implementation of this feature inside Renaissance is under our own
control, so if we decide to keep it in Renaissance, it is invulnerable to
change :-) it will stay there forever.


> and (and this is the main reason) it's also a bit awkward to handle.

Yes - I agree with that.  I don't particularly like it, yet it can be
handy if you call the method directly.

The reason I did it that way is that I wanted to have the .gsmarkup
loading API as similar as possible to the .nib loading API.  First, to
ease transition.  Second, because the .nib loading API has a very long
story and tradition and I assume it's a very good, mature starting point.

It's probably not a problem to extend it a little, if we do it properly
starting from practical problems of the existing one, and think carefully
about what we are doing. :-)