[Java-collab] Re: invokeOperation and Jody's stupid idea

[Jody Garnett]

> About this invokeOperation idea:
>
>> From a programmer's POV this is a great idea, since it means that  
>> it is possible to add new operations without touching the geometry  
>> interface, which I consider a good design choice.
>
> On the downside is that for the user buffer() or union() methods are  
> more intuitive than an invokeOperation(), but this is imho acceptable.

I think for common operations we can afford to break out methods that  
call invokeOperation; ie a buffer() and union() method would be good  
from a usability point of view. Indeed it would be nice to leave  
invokeOperation as an implementation detail; or an advanced method.
>
> P.S: The operations could be enumerations as well:   
> mygeom.invokeOperation(Operation.Buffer);

I kind of see that - you are declaring an option "buffer". And  
selection of an implementation (if indeed their is even an  
implementation is left to the the invokeOperation method to figure  
out. It can select one information for flat world; another  
implementation that does buffer on a sphere etc...
>
> That's it for now. I've got to catch a train and will be hiking the  
> next two days. :-)

Enjoy; bring lots of water.

Jody