Allow me to offer a suggestion regarding the meaning of "proxylet".
It has been common practice to use "X-let" to mean "a small portable program
that runs on an X". The problem with this approach is that is attempting to
characterize the program by the environment it executes in, and this *is
not* an inherent property of the program. Using this terminology, if I take
a servelet and run it on a proxy, it becomes a proxylet.
However, programs that run in a particular environment must be written to
the API that is supported in that environment, and the API *is* an inherent
property of the program. A program written to a servelet API is a servelet,
regardless of where I run it.
Thus, my proposal is that "X-let" be taken to mean "a small portable program
written to the X API".
Those of you how were at the NYC OPES workshop in Feb will not be shocked to
learn that I in fact intend to propose running servelets on proxy
intermediaries. Under this terminology, they do not thus become proxylets
because they do not make use of the proxy API, and so have no access to the