While reading this paper I had the impression that some good
thinking had gone into it. Unfortunately, I found it rather difficult
to follow partly because of the relatively inadequate English but,
perhaps more importantly, the vague and highly qualitative nature of
the arguments put forward. Instead of an explicit definition of
concepts we are given parables (e.g., the animals in the zoo analogy)
and examples from which we are asked to understand the proposed
concepts. (The examples are useful, but the explicit definitions are
still necessary.) I often found myself following and agreeing with
argumentation leading up to a certain conclusion and then finding
myself lost at the point when the conclusion is reached. That is, the
causal link between conclusion and argumentation was often unclear.
Also, I feel that the paper tries to say too much. There is the
concept of operations, which based on the informal description given
in the paper, does not seem to be distinguished significantly from
similar work. However, the paper then tries to tie this in with the
business of distribution configurations as well as the iterative
design process. The link between these things may very well exist (I
think so), but it is not quite clear from the paper what that link is.
My advice to the author is to break up the subject into smaller
paper-sized bites and present them in natural succession. For example,
the idea of operations (properly specified) probably warrants a paper
onto itself.
The author should probably look at the work of Nierstrasz
(Proceedings of OOPSLA'93) since this paper introduces the concept of
regular types that seems quite similar to "operations". Another useful
reference is the work by Yallin and Storm on protocols which also
define an asynchronous communication-based interfaces with a temporal
dimension.