Re: Graphics properties implementation

[Michael Goffioul]

On 1/17/08, John W. Eaton <address@hidden> wrote:
> So there would be a single validate function that switches on the
> property name?  That seems odd to me.  Why not individual functions
> like validate_faces (inserted automatically if the v qualifier is
> present)?  Then you would have to provide the validate_faces function
> by hand.

That's indeed a good solution as well (probably even better). My
intention was to trigger the discussion and try to get the best
solution out of it.

> Or, just provide the set_faces function by hand so that it
> can do the work directly.  Or is the problem with that approach that
> this kind of thing is needed for many properties?

Indeed, for complex objects like axes or patch, a bunch of properties
should trigger internal update. From experience, it's usually like
"a large set of properties trigger a small set of updaters". You can
have a feeling of it by looking at the "propertyChanged" methods in
http://octave.svn.sourceforge.net/viewvc/octave/trunk/octave-forge/extra/jhandles/src/org/octave/graphics/AxesObject.java?revision=4456&view=markup
or
http://octave.svn.sourceforge.net/viewvc/octave/trunk/octave-forge/extra/jhandles/src/org/octave/graphics/PatchObject.java?revision=4440&view=markup

> Is validate the right name?  Maybe "update_dependent_properites" (or
> something similar)?

Better, indeed. As I said, I wanted to trigger the discussion.

Michael.