Hi Boaz:

yes, a merge gets us close to the goal but there are two more issues:

1. The strategies define path sets, but we only need a strategy which
   defines a propagation graph, i.e., a subgraph of the original class graph.

2. There are many ways to condense the "big" merged strategy into a
   propagation graph defining strategy. To make marshaling and
   unmarshaling efficient,
   the condensed strategy should be of the form:
   "from A bypassing nodes and edges to *"
   For such strategies, the propagation graph contains loss-free
   information.

Which algorithm should we use to make the "big" merged strategy into
a simpler one for the current class graph?

-- Karl

From boaz@ccs.neu.edu Wed Oct  1 08:13:36 1997
From: Boaz Patt-Shamir <boaz@ccs.neu.edu>
MIME-Version: 1.0
To: Karl Lieberherr <lieber@ccs.neu.edu>
cc: dem@ccs.neu.edu, matta@ccs.neu.edu, boaz@ccs.neu.edu
Subject: condensing a set of strategies

Karl Lieberherr writes:
 > 
 > let's assume we have an object O on which we want to run in another
 > machine a set of traversals determined by a strategy set St.
 > What is the smallest subobject of O so that all those traversals
 > will run correctly? Can we efficiently condense the 
 > strategy set St into a strategy
 > of the form "from A bypassing ... to *" which can be used for 
 > marshalling purposes? 
 > 
Sounds to me like a simple merge: take the union of the constituent
strategies to obtain a multi-source, multi-target strategy which
covers precisely what you need.

Boaz