まさおのChangeLogメモ / 2005-03-15

- Flickr Postcard Browser: 402 users
- Open Clip Art Library :: openclipart.org: 6116 users
- CiNii :: Researchers Link Viewer: users
- FileSum ディレクトリの容量表示 FileSum: users
- ちらりウィンドウ（マウス用）: users
- akJ SubPlayer 字幕再生・MP3書き出し: users
- Webservices Mashup Resources - O'Reilly Etech 2005: 58 users
- ‘グリーン･チャリ東京’: users

最近、WinSCPを「最近使ったディレクトリに自動で移動する」のモードで
使っていて気付いたのだが、どうやら既に無いディレクトリが指定された
状態でログインしようとすると、そのままWinSCPはサーバからの返答が無
いとか言って、落ちるみたい（ログ参照）。

これは、ホームディレクトリにfallbackしてほしいなあ。
あとで要望を出しておこう。

ログ:
. 2005-03-15 16:56:25.469 --------------------------------------------------------------------------
. 2005-03-15 16:56:25.469 WinSCP Version 3.7.4 (Build 271) (OS 5.1.2600 Service Pack 2)
. 2005-03-15 16:56:25.469 Login time: 2005年3月15日 16:56:25
. 2005-03-15 16:56:25.469 --------------------------------------------------------------------------
. 2005-03-15 16:56:25.469 Session name: masao@nile.slis.tsukuba.ac.jp
. 2005-03-15 16:56:25.469 Host name: nile.slis.tsukuba.ac.jp (Port: 22)
. 2005-03-15 16:56:25.479 User name: masao (Password: No, Key file: No)
. 2005-03-15 16:56:25.479 Transfer Protocol: SFTP (SCP)
. 2005-03-15 16:56:25.479 SSH protocol version: 2; Compression: No
. 2005-03-15 16:56:25.479 Agent forwarding: No; TIS/CryptoCard: No; KI: Yes; GSSAPI: No
. 2005-03-15 16:56:25.479 Ciphers: aes,blowfish,3des,WARN,des; Ssh2DES: No
. 2005-03-15 16:56:25.479 Ping type: -, Ping interval: 30 sec; Timeout: 15 sec
. 2005-03-15 16:56:25.479 SSH Bugs: -,-,-,-,-,-,-,-
. 2005-03-15 16:56:25.479 SFTP Bugs: -,-,-
. 2005-03-15 16:56:25.479 Proxy: none
. 2005-03-15 16:56:25.479 Return code variable: Autodetect; Lookup user groups: Yes
. 2005-03-15 16:56:25.479 Shell: default, EOL: 0
. 2005-03-15 16:56:25.479 Local directory: C:\Documents and Settings\masao\デスクトップ, Remote directory: /tmp/mnewsprint-25665, Update: Yes, Cache: Yes
. 2005-03-15 16:56:25.479 Cache directory changes: Yes, Permanent: Yes
. 2005-03-15 16:56:25.479 Clear aliases: Yes, Unset nat.vars: Yes, Resolve symlinks: Yes
. 2005-03-15 16:56:25.479 Alias LS: No, Ign LS warn: Yes, Scp1 Comp: No
. 2005-03-15 16:56:25.479 --------------------------------------------------------------------------
. 2005-03-15 16:56:25.479 Looking up host "nile.slis.tsukuba.ac.jp"
. 2005-03-15 16:56:25.660 Connecting to 133.51.14.8 port 22
. 2005-03-15 16:56:25.700 Server version: SSH-1.5-1.2.31
. 2005-03-15 16:56:25.700 We claim version: SSH-1.5-WinSCP_release_3.7.4
. 2005-03-15 16:56:25.700 Using SSH protocol version 1
. 2005-03-15 16:56:25.710 Received public keys
. 2005-03-15 16:56:25.710 Host key fingerprint is:
. 2005-03-15 16:56:25.710 1024 78:92:60:c6:66:f1:ec:13:93:28:ee:ef:46:23:d1:7c
. 2005-03-15 16:56:25.710 Encrypted session key
. 2005-03-15 16:56:25.710 AES not supported in SSH1, skipping
. 2005-03-15 16:56:25.710 Using Blowfish encryption
. 2005-03-15 16:56:25.710 Trying to enable encryption...
. 2005-03-15 16:56:25.710 Initialised Blowfish encryption
. 2005-03-15 16:56:25.710 Installing CRC compensation attack detector
. 2005-03-15 16:56:26.280 Successfully started encryption
. 2005-03-15 16:56:26.280 Sent username "masao"
. 2005-03-15 16:56:26.290 Session password prompt (masao@nile.slis.tsukuba.ac.jp's password: )
. 2005-03-15 16:56:26.290 Asking user for password.
. 2005-03-15 16:56:28.223 Sending password with camouflage packets
. 2005-03-15 16:56:28.223 Sent password
. 2005-03-15 16:56:28.243 Authentication successful
. 2005-03-15 16:56:28.243 Started session
. 2005-03-15 16:56:28.243 --------------------------------------------------------------------------
. 2005-03-15 16:56:28.243 Using SCP protocol.
. 2005-03-15 16:56:28.243 Doing startup conversation with host.
. 2005-03-15 16:56:28.243 Skipping host startup message (if any).
> 2005-03-15 16:56:28.243 echo "WinSCP: this is end-of-file:0"
< 2005-03-15 16:56:28.313 Warning: no access to tty (Bad file number).
< 2005-03-15 16:56:28.313 Thus no job control in this shell.
< 2005-03-15 16:56:28.383 Sun Microsystems Inc. SunOS 5.7 Generic October 1998
! 2005-03-15 16:56:29.966 stty: : Invalid argument
< 2005-03-15 16:56:30.186 WinSCP: this is end-of-file:0
. 2005-03-15 16:56:30.186 Detecting variable containing return code of last command.
. 2005-03-15 16:56:30.186 Trying "$status".
> 2005-03-15 16:56:30.186 echo "$status" ; echo "WinSCP: this is end-of-file:0"
< 2005-03-15 16:56:30.316 0
< 2005-03-15 16:56:30.316 WinSCP: this is end-of-file:0
. 2005-03-15 16:56:30.316 Return code variable "$status" selected.
. 2005-03-15 16:56:30.316 Clearing all aliases.
> 2005-03-15 16:56:30.316 unalias "echo" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:30.456 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:30.456 unalias "pwd" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:30.587 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:30.587 unalias "cd" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:30.707 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:30.707 unalias "ls" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:30.837 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:30.837 unalias "groups" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:30.967 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:30.967 unalias "scp" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:31.107 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:31.107 unalias "rm" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:31.218 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:31.218 unalias "mv" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:31.338 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:31.338 unalias "mkdir" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:31.468 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:31.468 unalias "chmod" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:31.598 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:31.598 unalias "chgrp" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:31.728 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:31.728 unalias "chown" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:31.848 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:31.848 unalias "unset" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:31.979 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:31.979 unalias "unalias" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:32.099 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:32.099 unalias "alias" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:32.219 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:32.219 unalias "ln" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:32.339 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:32.339 unalias "cp" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:32.469 WinSCP: this is end-of-file:0
. 2005-03-15 16:56:32.469 Clearing national user variables.
> 2005-03-15 16:56:32.469 unset "LANG" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:32.600 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:32.600 unset "LANGUAGE" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:32.730 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:32.730 unset "LC_CTYPE" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:32.860 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:32.860 unset "LC_COLLATE" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:32.990 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:32.990 unset "LC_MONETARY" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:33.120 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:33.120 unset "LC_NUMERIC" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:33.250 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:33.250 unset "LC_TIME" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:33.391 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:33.391 unset "LC_MESSAGES" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:33.521 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:33.521 unset "LC_ALL" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:33.661 WinSCP: this is end-of-file:0
> 2005-03-15 16:56:33.661 unset "HUMAN_BLOCKS" ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:33.791 WinSCP: this is end-of-file:0
. 2005-03-15 16:56:33.791 Looking up groups and users.
> 2005-03-15 16:56:33.791 groups ; echo "WinSCP: this is end-of-file:$status"
< 2005-03-15 16:56:33.951 ulis ntcir nas2003 goi ethome
< 2005-03-15 16:56:33.951 WinSCP: this is end-of-file:0
. 2005-03-15 16:56:33.951 Following groups found:
. 2005-03-15 16:56:33.951 ulis
. 2005-03-15 16:56:33.951 ntcir
. 2005-03-15 16:56:33.951 nas2003
. 2005-03-15 16:56:33.951 goi
. 2005-03-15 16:56:33.951 ethome
. 2005-03-15 16:56:33.951 No users found.
. 2005-03-15 16:56:33.951 Changing directory to "/tmp/mnewsprint-25665".
> 2005-03-15 16:56:33.951 cd "/tmp/mnewsprint-25665" ; echo "WinSCP: this is end-of-file:$status"
! 2005-03-15 16:56:34.072 /tmp/mnewsprint-25665: No such file or directory.
. 2005-03-15 16:56:49.073 Waiting for data timed out, asking user what to do.
. 2005-03-15 16:56:49.073 Asking user:
. 2005-03-15 16:56:49.073 Host has not answered for 15 seconds.
. 2005-03-15 16:56:49.073
. 2005-03-15 16:56:49.073 Wait for another 15 seconds? Pressing 'Abort' button will close session. ()
. 2005-03-15 16:56:51.707 Attempt to close connection due to fatal exception:
* 2005-03-15 16:56:51.707 Terminated by user.
. 2005-03-15 16:56:51.707 Closing connection.
* 2005-03-15 16:56:51.727 (ESshFatal) Error changing directory to '/tmp/mnewsprint-25665'.
* 2005-03-15 16:56:51.727 Terminated by user.

doi:10.1023/B:INRT.0000011206.23588.ab
TRECおよびSPIRITコレクションを使って、よりリアルなWebコレクション
の形成に必要なリンク密度について解析。

・書誌情報:
Cathal Gurrin, Alan F. Smeaton:
Replicating Web Structure in Small-Scale Test Collections,
Information Retrieval, 7(3-4), pp.239-263, 2004.

・内容:
Web上のリンク解析ではサイト間のインリンクが重要で、この密度が低い
と思ったような結果が得られない。TREC8〜TREC2003までの結果を見ると、
WEB-IRでリンク解析を用いたシステムは低い結果しか得られていない。こ
れは現在のテストコレクションのリンク密度がかなり低かったためだろう。
そこで、リアルなWebに近いリンク密度を達成する必要がある。
Web上のリンク密度は、ほぼ以下の通り:
・サイト内リンク: 一文書平均 14.2
・サイト間リンク: 一文書平均 4.9
さらにリンク数の分布はpower-lawに従う。
これらの条件に合うようなテストコレクションを構築手法について考察。

・総評（感想）:
前半のリンク情報の解析や必要なリンク密度に至る実験やデータの説明は
比較的分かりやすく、論文も読みやすいが、肝心の「どうやってうまくテ
ストコレクションを構築していけば良いか」についての考察が薄い。

NW1000G-04についても、この論文で示されたデータに沿うリンク密度になっ
ているか、要考察。再現性を実験すること。

an overview of a large web collection:
doi:10.1145/1041394.1041395
テラバイト級のテストコレクションSPIRITについての統計情報について。
SIGIRのニュースレターに載った上保の論文、というか技術報告。

・書誌情報:
Hideo Joho, Mark Sanderson:
The SPIRIT collection: an overview of a large web collection:
ACM SIGIR Forum, 38(2), pp.57-61, 2004.

・概要:
SPIRITはEUのプログラムにより構築されたほぼ1TBのWEBテストコレクショ
ンで、地理情報系のIRが主な目的。元々はUniv.of Waterlooでクロールし
たデータの一部から作成されたもの。

以下の統計情報について記載:

・文書数
・文書数（サイト単位の平均/最大）
・サイズ
・サイズの分布（{10,20,50,100,1k,2k,5k,10k,20k,...1m,2m,5m,10m,...）
・表示ワード数の分布（lynxのダンプ）
・ドメイン分布（トップレベル・2ndレベル）
・国
・charset（<meta>内）
・URLの深さ（/の数）
・IMGタグの数（ALT有/無）

リンク解析については別論文を参照している。