# [e-lang] Quick comments needed on tgc draft

David Hopwood david.nospam.hopwood at blueyonder.co.uk
Tue Jun 21 14:49:17 EDT 2005

Mark Miller wrote:
> David Hopwood wrote:
>
>>>> Terminology issue: do we count systems with multicast or other
>>>> multiple-target messaging primitives as message-passing systems, or
>>>> is single-target implied unless otherwise specified?
>>>
>>> I would count them as message passing. Joule's Channels are
>>> multicast, and I'd certainly count Joule as a message passing system.
>>> But I doubt this distinction is relevant within the scope of this paper.
>>
>> OK, then add a footnote "We only consider messages sent to a single
>> object, although some message passing systems support multicast messages."
>
> Is there some potential point of confusion you're worried about here?
> (If there is, I'm happy to include the footnote.)

I suppose I'm just being excessively pedantic. The original caption sounded
as though it was referring to all message passing systems. Maybe it was only
referring to E.

>> I've added macros to the LaTeX document (attached) so that the styles can
>> be easily changed:
>> [...]
>> I've also changed the code samples to use the alltt environment instead
>> of verbatim, to allow styles to be used for method and variable
>> declarations.

The styles that used \emph (italics) should probably be slanted instead:

\newcommand{\name}[1]{{\sf\textsl{#1}}}        % objects/processes
[...]
\newcommand{\var}[1]{{\tt\textsl{#1}}}         % variables
\newcommand{\dvar}[1]{\textsl{#1}}             % declarations of variables
\newcommand{\dobj}[1]{\textsl{#1}}             % declarations of objects

> Fortunately, since I spent the day futzing around (badly and stupidly)
> with bibtex, our changes were easy to integrate.

The LLNCS style apparently wants article titles to be mostly lowercase, but
is sometimes quite overzealous at it. For example:

M. Allman. An evaluation of xml-rpc.

This can be fixed by adding curlies around acronyms, like so:

title = {An evaluation of {XML-RPC}},

> The integrated doc can
> be found at the old URL
>
> https://www.cypherpunks.to/erights/talks/promises/tgc05.tex
> https://www.cypherpunks.to/erights/talks/promises/tgc05.pdf
>
> I've also posted
>
> https://www.cypherpunks.to/erights/talks/promises/stackvat.svg
> https://www.cypherpunks.to/erights/talks/promises/stackvat.eps
> https://www.cypherpunks.to/erights/talks/promises/common.bib

stackvat.eps isn't linked from the HTML index. Also, a .zip or .tgz of all the

> The stackvat files are David's variation on Ping's drawing. For the
> common.bib file, I started from a file I got from Matej, got rid of much
> of the stuff I don't expect to cite, and did various web searches to
> find bibtex entries for the stuff I want to cite. However, a lot of
> these are wrong, and I haven't fixed them yet. Where do people generally
> go to find correct bibtex data?

ACM and Springer have accurate bibtex entries for their papers. Otherwise citeseer,
but if there's any reason to doubt the accuracy of that data, I download several
citing papers and merge the information in them.

--
David Hopwood <david.nospam.hopwood at blueyonder.co.uk>