Re: computer-go: Proposal for an electronic journal of computer go

[Martin Mueller <mueller@xxxxxxxxx>]

Hello all,

I have been traveling for a while, but I have now caught up with my email
and read most of this thread, which is of great interest for me. I'll try
to add some personal viewpoints to the discussion.

First of all, I think an electronic computer Go journal could be a
wonderful thing. But as others have pointed out, there are already a number
of alternatives, and there is the big problem of finding volunteers to do
the actual work. For the old Computer Go journal the one person doing
almost all the work was the editor, Dave Erbach. That journal stopped
publication when Dave became the department chair at his university, which
left him with no spare time for the editing work. Unfortunately, history is
repeating itself to some degree with the AGA computer Go site, for which I
created the content. In the last year or so, I have been so busy that I did
not find the time for creating an update, although I do keep collecting
interesting stuff. My efforts to find someone to help me with the editing
and preparation for web publication have been frustrating and ultimately
unsuccessful. For the AGA site as well as for a possible electronic
journal, we need volunteers that can work reliably over an extended period
of time, say at least 3 years. This is by far the most important point in
my opinion.

Regarding the scope of a journal, the old CGJ contained community service
sections such as announcements and tournament reports as well as articles
about Go programs, techniques, etc. The service part has been taken over to
some degree by the AGA and many other web sites. Some of this information
also goes into the ICGA (formerly ICCA) journal. As for papers, Go
publications have occasionally found their way into the mainstream AI and
CS publications, and more are published in games-specific conferences and
workshops such as IJCAI'97 workshop, CG'98, and the GPW series here in
Japan. So one strategy is simply to try to increase the number of Go papers
in these refereed publications. The next big conference will be CG'2000,
with a deadline for submissions in May. See links below. A good Go paper
should have no problem getting into one of these conference proceedings,
even if the work is very Go-specific. Furthermore, I think it's a good idea
to regard research in computer Go not as a goal all by itself, but as a
contribution in the larger context of games, AI, knowledge engineering, and
search. Writing papers from that perspective definitely helps to get them
published.

Regarding the (unrefereed) preprint server that Dan Bump mentioned, I think
it already exists in the shape of Markus Enzenberger's excellent Online
Computer Go Bibliography. Of course, it may be useful to collect copies of
files (with the author's permission) in one archive to prevent link
failures.

One important niche that is not covered by any of the existing mechanisms
(as far as I can see) was mentioned by Michael Collins:

>     ... I'd like to see if we can include information on the strengths
>and weaknesses of current go software, possibly .sgf format files of games
>from the computer go ladder etc.
>     It would be nice to see papers and articles on approaches to modelling
>go, possibly seeing if authors would contribute their own approaches as
>well.

Yes, definitely. There are many Go programmers that are afraid to write for
an academic type of journal but might be cinvinced to publish in a more
informal way. Also, it would be great to have e.g. a Java viewer for
commented computer games for the collection at AGA. My colleague Reijer
Grimbergen has such a site for (human) shogi game records (see link).

To sum up, I support the idea of reviving the Computer Go journal in an
electronic form, but I also wonder if it wouldn't work better to just try
to improve and better use existing resources, and maybe bring them all
together under one roof such as a computergo.org website.

	Martin

Links:

http://www.etl.go.jp/etl/suiron/~ianf/cg2000/index.html
http://home.t-online.de/home/markus.enzenberger/compgo_biblio.html
http://www.usgo.org/computer/
http://www.dcs.qmw.ac.uk/~icca/journal.htm
http://www.etl.go.jp/etl/suiron/~grimberg/SHOGI/kifu.html