Thank you, Martin! May you also provide the metamodel as .ecore file? Dot-ecore is the EMF fileformat, as I understand, right? Also I saw you have association subclasses for class cast and instance of checks. I like that approach, do you also have done or thought about extending the model for annotations and generics? cheers, AA On 28 Jun 2007, at 11:52 , Martin Pinzger wrote:

Hi, That is nice, because we ended up with a similar solution, but instead of an integer we use a string. When the data is unknown, we do not specify it. Doru On Jun 28, 2007, at 5:16 PM, Stéphane Ducasse wrote: -- www.iam.unibe.ch/~girba www.iam.unibe.ch/~girba/blog/ "In a world where everything is moving ever faster, one might have better chances to win by moving slower."

Hi Martin, I have some questions/remarks regarding the picture: - What does Association.statement point to? - Is there any particular purpose for the Context class? - Should File and Directory have modifiers? - It seems to me that to model the Associations uniformly, you deliberately do not use real static typing for from and to :). The same goes for parent. Am I right? - Instead of parent, I propose container, because parent is overloaded with respect to inheritance (e.g., parent class). - Instead of from and to, would you mind having full names (e.g., sender, accessor) or do you have a particular reason for having them like this? Cheers, Doru On Jun 28, 2007, at 7:14 PM, Adrian Kuhn wrote: -- www.iam.unibe.ch/~girba www.iam.unibe.ch/~girba/blog/ "Problem solving efficiency grows with the abstractness level of problem understanding."