新城です。

In article <030819001919.M0102865@ims.ipc.kit.ac.jp>
        chiaki@ipc.kit.ac.jp (Tsukamoto Chiaki) writes:
> 工繊大の塚本です.
> % From: "KalElFan" <KalElFan@scifipi.com>
> % Newsgroups: news.groups
> % Subject: "Optional" Moderation
> % Message-ID: <adB_a.8328$Z03.605162@news20.bellglobal.com>
> % Date: Wed, 13 Aug 2003 21:03:51 -0400
> 
> から始まる thread で, 未だ続いているのではないかしらん. この
> 記事は読む意味は余りないでしょうから, 次のものを御参考までに.

あ、やば。うちのサーバに news.group の記事が全然届いていませ
ん。よく見ると、クロスポストされたものだけだ。

> From: Russ Allbery <rra@stanford.edu>
> Newsgroups: news.groups
> Subject: Re: "Optional" Moderation
> Date: Wed, 13 Aug 2003 20:24:18 -0700
> Message-ID: <87n0ecc465.fsf@windlord.stanford.edu>
> This was exactly the problem that NoCeM was designed to solve.  It didn't
> catch on in part because implementing this securely proved to be somewhat
> difficult, but more importantly because everyone wants to use a system
> like this and no one wants to be the moderator.  This is the same problem
> that the GroupLens project ran into:  everyone wanted to use the ratings,
> but no one wanted to do the work to submit their own ratings.

みんな moderator が厳選した記事を読みたいけれど、誰も
moderator にはなりたくない。
GroupLens って何なんでしょうね。

> After having watched a few projects like this start and fail, I don't
> think the basic theory is the missing piece.  I think the missing piece is
> some way of making the communication between the moderators and the users
> very lightweight, simple, easy to understand, and exceptionally easy for
> the moderators to generate (ideally, nearly automatic).  That's the place
> that really needs a lot of work, and I personally don't have any brilliant
> ideas about how to get there.

足りないものは、理論じゃなくて、moderator とユーザの間のやり
とりを軽く、理解しやすくすること。moderator には格別に簡単に。
やるべきことはあるけど、まだいいアイディアは浮かばない。

fj.news.readers の話も、こんな感じ。そんなに負けていませんね。
fj の場合、状況が少し違うのは、どちらかというと記事を増やす
という話も重要だということでしょうか。

> I think these sorts of systems have difficulty in part because they have a
> built-in freeloader problem.  It's a *lot* of work to do the moderation,
> lots of users just benefit from it without having to do anything at all,
> and it's very difficult to share the work in any fashion that doesn't end
> up putting all the burden on a few people who get burned out and quit.
> Regular moderation has this problem, and trying to extend that work to a
> substantial number of unmoderated groups is far more work than is already
> going into the maintenance of the existing moderated groups.  Many of
> which, recall, are dead because people burned out on being moderators.

moderator が便利なのはみんな分かっているんだけど、たいへんな
ので燃え尽きちゃう。

> People seem to badly underestimate the amount of work that goes into
> moderating a group well enough for the results to be useful.  It's not
> like spam filtering.  Usenet has almost no spam in the e-mail sense and
> the spam that it does have is fairly easy to control if you want to do so
> (there are lots of spam filters for news servers, people who issue
> cancels, and so forth).  Usenet has a harder problem, namely off-topic,
> inflammatory, or just plain unuseful postings, and these are difficult to
> classify, controversial, and require a lot of human work to filter out.

spam を弾くのは、比較的簡単だ。残っている難しい問題は、しょ
うもない記事を弾いたりすることで、それには人手がかかる。

> > >  vote (dvote) という研究は Russ Allbery さんの所で行われて
> > > いたのでありました.
> 実験者の意識が NetNews の方よりも, Web に参加者が有用な link
> を持ち寄ったり, 掲示板サイトとして参加者同士会話したりして,
> 共同体を作り上げる方に向いていたので, NetNews への批評サイト
> としては十分に機能しなかったと言えるかも知れません.

内向きの共同体か。fj.net.watch は、共同体になるのかなあ。

fj の問題は、USENET とはちょっと違うかもね。記事が少ないと 
rating のしようもないから、まず記事を増やすというのが fj の
課題なので。

＼＼　新城　靖　（しんじょう　やすし）　＼＼
＼＼　筑波大学　電子・情報　　　　　　　＼＼