On 15-Apr-08, at 8:46 PM, Colin Putney wrote:
>
>
> The usual way to do this is to have all your undoable units of work  
> be implemented by "command" objects that know how to perform a  
> specific action. Then when the user does something undoable, you  
> create a command, execute it, then push it on the undo stack. To  
> undo you can either have the command know how to undo its self, or  
> have it create an "inverse" command. This kind of thing works pretty  
> well, and it can be extended with other features like redo, logging,  
> journal-and-snapshot persistence etc.
That's pretty much what we did in Sophie. It's non-trivial but then  
the problem domain is non-trivial.

tim
--
tim Rowledge; tim at rowledge.org; http://www.rowledge.org/tim
A sad tale that brings a lump to the eye and a tear to the throat.