Hi Johan:

>From johan@ccs.neu.edu Thu Jan 14 17:32:38 1999
>To: Karl Lieberherr <lieber@ccs.neu.edu>
>Subject: Re: TAO etc. 
>From: Johan Ovlinger <johan@ccs.neu.edu>
>
> Oh, I remember now: you have line strategies. Why does your algorithm handle
> short-cuts correctly? They already happen with line strategies.
>
>short cuts? hrm. yes, I do believe that I get those right. The problem
>only arises when we try to encode A . B into one subgraph object. I
>keep separate subgraph objects for A and B. (A,B are strategies, and a
>subgraph object is the path of the class graph that is to be traversed
>by a strategy).

That sounds right. So TAO is correct but works only for a limited set
of traversal strategies.

>
> You mentioned that you alredy have some code in this direction.
>
>I think you're misremembering. I do have TAO, but that is it. However,
>it should be fairly quick to get it back into shape, I think. 
>
>Johan
>

Once the reflection solution works, I would like to reactivate TAO
as a part of Demeter/J through a student project with you as 
consultant.

Please can you write the spec for the student project.

-- Karl