home > plop > 2006 > Papers > Notes > 44-notes.php

PLoP 2006 Shepherded Papers

Notes on Paper 44:

This paper's technical level and list of references are impressive, but
it is not very well written patten paper. It uses mainly GOF-format,
but curiosly enough add forces section to it. It is very intersting
that in pattern writing point of view this forces section is better
than many papers usually has. Unfortunately most of the other "pattern
things" are mixed in motivation section.

I maybe make wrong recommendation to them to propose them to start to
change this more "Coplien format paper", because most of the
problem-context is already there somewhere. Hidden and mixed, but
existing. There are significant unbalance between "implementation part
of paper" and "pattern part of paper".

Anyway, they change paper during shepherding only very little and we
are not able to complete more than one iteration. Part of the reason
they gave to me is that two of the authors are not reachable by email
because of either vacation or middle-east conflict (they are in
Israel).

When reading this paper (and especially references section) I have
feeling that authors have lot of work in this technical area, but not
much knowledge about patterns. The refer only to GOF, not any other
patterns which have similarities to their pattern e.g. Half-Object.

 From pp. 2 "The main idea of the pattern is to share the same stub
code between client and server applications by the use of a common base
class, instead of generating the stubs from the same interface
definition file."

Also, this paper is (see quotation below) either usage of GOF-patterns
or modification of them.

 From pp.2 "The RPC Arranger pattern borrows its structure from the
known Bridge and Template design patterns 14, and the Dispatcher design
pattern [GOF]."

So, my recommendation. I am in the middle of go/no-go. We do not have
successfull shepherding (partly my fault, probably) and there are good
parts of paper. But I have some suspicious that authors do not get
benefit from workshop. Also, I cannot help myself have feeling that
this paper's technical part have been in other kind of conference
submission and pattern part is added later for PLoP submission

--
Juha Pärssinen, juha.parssinen@vtt.fi