source: Papers/fopara2013/appendix.tex @ 3452

\section{Changes in response to the reviewer's comments}

\begin{itemize}
\item
We have made more explicit reference to a key WCET survey paper, in
particular noting that it is a useful source of information about
commercial tools such as AbsInt.

\item We have altered and improved the example to remove some
distracting ambiguity and demonstrate a parametric time bound.  The
text now identifies the added code by more than just its colour.  The
later sections now refer back to the example to illustrate the effects
of each technique.

\item
We've removed the paragraph on logic programming languages because one
of the reviewers pointed out that standard Prolog may have the
necessary features, which we haven't tried.  Thus we felt it was best
not to claim anything about logic programming at this time.

\item
We have been more explicit about the type of code generated by the
Lustre compiler, and what kinds of loop can be handled automatically
by the Cost plugin and where loop iteration bounds come from.

\item The structure of the document has been greatly improved.
Section 2 has been restructured and broken up into subsections to give
a clearer comparison of existing approaches with CerCo.
Section 4 has been reorganised to highlight the connections between
the different parts of the work and a short overview provided to help
the reader understand the significance of each part of the work.

\item A clearer distinction is made between the core contribution of
the compiler --- which provides local cost information in the source
code and thus is largely agnostic to how loops are structured and what
kind of data structures are used --- and the cost synthesis work which
provides higher-level costs using it.

\item The text has been extensively revised to be easier to read, and
to make minor corrections, including those identified by the reviewers.

\end{itemize}