17)     The document addresses the usage of raw compute facilities - processors, storage,  network - when running a "job".  Would it make sense to address other  cases, such as tracking the usage of a particular application?  This is important  e.g. for licensing models - see the discussion panel on licensing at GridWorld  2005.

I see two sub-cases.  In the first, the computing resource provides the application.   In the second, it is staged in as part of the "job".  These may  need to be addressed differently.

Evaluation: This is definitely an important area, but out of scope for the current document. The topic will be taken to the mailing list for discussion.