Here are some correction/remarks I proposes.

1 Introduction
---------------
- Last pararaphe on page 3, before section 1.1. "...influence the design of a version 2.0 of SAGA." should be replaced 
by "...influence the design of future versions of SAGA." We should not limite the scope to version 2.0 solely. page 3.

2.2 Functional Areas
------------------------
Table 1 should be put after the enumeration not within. page 9.

2.3 Non-Functional Areas
------------------------
Table 2 should be put after the enumeration not within. page 13.

Quality of Service: should this not be better to mention "...This issue should be revisited for future SAGA versions." 
instead of "... This issue should be revisited for SAGA version 2.0." page 15.

4.3 Language bindings
--------------------------
There is no explicite reference to the L&F of the API itself. I mean the naming convetion, OO design (patterns, 
factories, etc.) , packaging, namespace, etc. Only about error handeling... If this is a binding issue then it should be
 clearly stated in the document. 

My concern is that we will see a pseudo C implementation in Java, or something similar. If that happens the Java 
community will feel that this API is not a Java binding but some C wrapper (see the corba for Java implementation). 
Similar for other languages. I mention Java because I'm used to that language and the L&F is pretty well defined in the 
SUN documentation even if often not respected.

(I didn't check the OMII implementation yet, so I might be wrong, in that case please ignore this comment)


References
-------------
- Item [8], "Naregi, Japan..." should be "NAREGI, Japan. The National Grid Research Initiative project". page 29


Otherwise I think this document is good. Excellent job Shantenu and Andre (and Tom)!