[aosd-discuss] Java project to AspectJ

[Raffi Khatchadourian]

David,

david shepherd wrote:
>  Yes, exactly, this would be an EXCELLENT evaluation, but is very difficult to create this benchmark correctly.  
>   
These are valid points. In that case, a good benchmark would be a large 
open source project that been altered manually to use aspects. One 
project that contains aspects that may have not in the past is Eclipse. 
If you can get your hands on "pre-aspect" code, run your mining tool on 
this code, and then compare it to later code that contains aspects, you 
may come up with some insightful results. One unfortunate downfall of 
this approach is that in may be the case that more of the original code 
might have changed besides adding aspects.

Raffi