Cataloges:   Monographs || Serials || Videos || Sound || EVOCS || SSELP || DBs || DACHS || IGCS || Heidi || CrossAsia || KVK || ProQuest  
Ruprecht-Karls-Universitat Heidelberg

Using the Academia Sinica's Full Text Retrieval System (FTRS)
中文全文檢索系統


Top | Introduction | Telnet Version | Special Characters | Navigating the Database | Searching the Database | Further Information

1. Introduction

The Academia Sinica's Full-text databases are among the most advanced research tools available for Chinese Studies. They consist of very reliable and substantial electronic texts which also contain commentary and subcommentary where available, and an extremely powerful search engine which makes these texts accessible.

Two interfaces are available for these databases:

The powerful search engine is with all its advanced features is only available with the telnet version. Therefore, for serious scholarly research the telnet version is strongly to be recommended.
Some important advantages of the telnet version over the WWW version:

  • possibility to limit searches to certain sections of a text (e.g. exclude commentary from a search)
  • advanced options for result display (list form, with sorting options)
  • possibility to use previous search result set as search range for the next search
  • versatile navigation in the text


Top | Introduction | Telnet Version | Special Characters | Navigating the Database | Searching the Database | Further Information

2. Preparation for using the telnet version:

  1. Make sure your computer has a working connection to the Internet.

  2. Get a login name and password from library.uni-heidelberg.de/
    (access restricted to European users only; no login necessary from within the institute).

  3. Get a SSH compliant telnet client
    for Windows, SecureNetTerm can be recommended, which is available for download from http://www.netterm.com/html/securenetterm.html (SecureFTP.exe - Windows 95/98/NT/W2K/ME/XP 32 bit English version 6.6.0.2003).
    for Apple MacIntosh we recommend MacBlue Telnet, which can be downloaded at http://icm.allmacintosh.com/preview/206677.html (ver. 2.6.6 of Oct 5, 1999, 1.1MB).

  4. Login to 147.142.222.48, port 23, emulation VT 220 (this is the default, which can be changed), connection type TCP/IP (in Netterm, this is best done in the address book).

  5. Either use Traditional Chinese Windows 95, 98 or NT, or use Windows 2000 with support for Chinese installed, or get a Chinese environment and set it to Big5. You may need to switch on auto-mapping of English fonts to Chinese and to choose an appropriate font in the set-up of your telnet client (in SecureNetTerm, isi_ansi will usually do). Some Chinese environments are:
    (a) for DOS: Yitian, Guoqiao;
    (b) for Windows: Nanjistar Communicator, Twinbridge Chinese Partner, Richwin, Unionway Asian Suite, Winmass.

  6. Install the userfont:
    (a) For a DOS-based telnet client used with ETEN (Yitian), get the appropriate userfont files from library.uni-heidelberg.de.
    (b) For Traditional Chinese versions of Windows 3.1, Windows 95 or higher download the appropriate userfont files from http://sino-sv3.sino.uni-heidelberg.de/anyang/ufsetup.html or from http://www.sinica.edu.tw/~tdbproj/handy/ufsetup.html (the latter source has instructions for Windows 98) and follow the instructions given there for set-up.
    (c) If you need to use more than one userfont in Chinese Windows 95 or higher, contact luedke.uni-heidelberg.de for instructions.

    To our knowledge, there is no way to use the userfont files in non-Chinese Windows 3.1, 95, 98 or in a non-Windows operating system (except DOS). All language versions of Windows 2000 should be able to install the userfont files. If you have experience or need help with userfont files in Windows 2000, please also contact edv@uni-heidelberg.de.

  7. If you want to export data to your computer, you also need a ftp client (a graphical client is usually found easier to use). Export data, which is first saved on the server in Heidelberg, can be accessed at 147.142.222.48. After anonymous login with your login name and password, you should be in the appropriate directory where the export file is located (resource5/download25, mapped as root directory after anonymous login).

  8. Be aware that certain features (search and retrieval etc.) are provided by the database application and cannot be changed, while other features such as font and colour on your own computer screen, screen shots, logging, local printing and so on are determined by your telnet client and can be set-up by yourself. Advanced telnet clients, such as netterm, provide a script language to simplify the login procedure. A simple script to short-cut the login procedure can be obtained from edv@uni-heidelberg.de/.
    If you want to call up netterm from the command line or from within a batch file with these scripts, use these commands:

    c:\win95_tw\command\start d:\netterm\netterm.exe 25histories
    c:\win95_tw\command\start d:\netterm\netterm.exe 13classics


Top | Introduction | Telnet Version | Special Characters | Navigating the Database | Searching the Database | Further Information

3. Problems associated with characters not included in standard fonts

1. Non-standard characters

2. Alternative graphical forms of characters which are included in standard fonts

The complete database of the 25 Dynastic Histories contains for:
28.306 instances (96%) faithful representation of the printed version, not included in the standard font;
1.189 instances (4 %) the form that is included in the standard font;

13.129 instances faithful representation of the printed version, not included in the standard font;
群 190 instances (4 %) the form that is included in the standard font.

This poses problems for input, display as well as search strategies.


Top | Introduction | Telnet Version | Special Characters | Navigating the Database | Searching the Database | Further Information

4. Navigating the database

This section assumes that you have already logged into the database. For details on preparing your computer and logging in, please see above.

4.1 Basics

The database is controlled by menus and commands.
The screen is divided into a narrow top section and a broad bottom section.
The top section can either be one menu line or two path lines. A menu line consists of different menu options. Of the two path lines, the second consists of the path field on the left (路經 e.g. /2.1.1.8.2.1) and a command line for inputting path statements on the right (指令 ); the first path line above is a line in reverse colours which displays the path natural language (e.g. 新 本 史記 三 家注/ 新 本 史記/ 本 紀/ 券 八 高 祖 本 紀 第八).
The bottom section is either blank (or contains some system information, e. g. 費 時 0. 15 秒- 找到 9 段 , 9 , 12 詞), or it contains continuous text (i. e. the full text), or it contains a result list, or it contains a directory list.
Which commands are available depends on which of these four alternatives is displayed in the bottom section. No commands are available if the bottom section is blank. If a path line is displayed in the top section, it is always possible to input path statements.

4.2 Commands

To use commands, just press the corresponding key on your keyboard. Except when inputting into the command line, the cursor position does not matter, and you will not see your input on the screen.

4.2.1 Commands for navigating lists

browsing the list:

↓ or + move to next line below
<n> ↓or + move down <n> lines
↑or - move to next line above
<n> ↑ or - move up <n> lines
<n> l move to line number <n>
n or SPACE next screen
<n> n or SPACE move forward <n> screens
b back one screen
<n> b move back <n> screens
$ display very last screen
F define search term and find in this display
f continue with find

changing the directory/move to text section:
g go to the item where the cursor presently is (i.e. display a subdirectory or text section)
q or t exit (quit, terminate)

4.2.2 Additional commands for controlling directory lists

. move up one level
/ move to root directory
x move to first page of the current directory item
y move to last page of the current directory item
<n> p move to page number <n> - works only if the directory item has to the right either a * or a number (refers to total number of pages in that section); all page numbers refer to the printed text in the 中華 書 局 standard edition of the Dynstic Histories
: move the cursor to the command field (指令:) in the right upper corner, where commands to move to a certain path or page can be inputted (see below)

other commands

<n> d set directory to be displayed <n> levels deep (useful for orientation before setting search range or area)
c choose content type (e.g. headings, commentary, philological notes) - only the content type chosen will be displayed and searched; in the frame that appears, use the following commands: m (mark) to choose, q or t (quit, terminate) to save and exit. For further details, see below 5.4.
m mark or unmark a specific directory item as search range
M cancel all marked directory items

4.2.3 Commands for navigating continuous text

SPACE, RETURN display next screenful of text
b back to last screen displayed
] move to beginning of next section
[ move to beginning of last section
a set display properties
1. 依 系 統 設 定 use system defaults (horizontal, continuous word wrap)
2. 重 行 設 定 set new display properties
a)呈現 方' (0. 橫 1. 直) vertical (0) or horizontal (1) display
b)呈 現 方' (0. 橫 1. 直) continuous (y) or non-continuous (n) word-wrap: continuous word-wrap is easier to read, non-continuous word wrap results in a layout that corresponds to the printed version, i. e. every line begins with the same character as in the printed version, making it very easy to locate a certain passage
T. 離開 exit without change
e cut commands in order to cut out text which can be used for further search statements: c defines cursor point as starting point of cut operation; then move cursor to end point and press c again
b      check the text which has been cut out
d      cancel all text that has been cut out
q, t    exit the cut operation
g move to the next text section in the search result set
G move to the previous text section in the search result set
p display text page by page (beginning of screen corresponds to beginning of page in printed edition)
s display text section by section (beginning of screen corresponds to beginning of section in the database's own segmentation, as indicated by the path information (lu4jing4, top left corner of screen, see 4.1 and 4.2.4)

4.2.4 Commands for path statements and page

For basic information on the path structure see 4.1 above.

Path commands refer to the present path if not preceded by / .

/ (slash) move to root path
. (dot) move up one level in the path hierarchy (each . moves up one level, more than one . moves up the corresponding number of levels)
p move to page number in the printed version (華書局 edition for 二十五史; 藝文印書館 edition for 十三經注疏 ), if the page number is unambiguous - e.g. /2 refers to 新 本 史記 三 家 注, but /2.20p is an invalid command, as there are three sections (新 本 史記 , 三 家 注 序 , 點 後 記), each with its own page numbering starting from page 1).
x moves to the first page of the first section of the cursor item
y moves to the first page of the last section of the cursor item
u defines the present path (see path field) as root

path and page commands (not x, y, u) can be combined
Example: Present directory path is /3.2, then the following applies:

Command Result
4 /3.2.4
...2.1 /2.1
x first text page of section /3.2
55p /3.2.55p, i. e. page 55 in the main text of the Hanshu
/2.1.120p /2.1.120p, i. e. page 120 in the main text of the Shiji


Top | Introduction | Telnet Version | Special Characters | Navigating the Database | Searching the Database | Further Information

5. Searching the database

5.1 Selecting a database (25 Histories 二十五史 or 13 Classics 十三經 )

For all work with the databases, it is necessary to first open a database by selecting from the menu 0) 资籵庫 and then to select one of the choices (at present 二十五史 or 十三經). Confirm by pressing the "Enter" key.

The retrieval system available at Heidelberg contains at present:
a) the 25 Dynastic Histories with commentaries, based on the standard Beijing Zhonghua Shuju edition of the Dynastic Histories.
b) the 13 Classics with commentary and subcommentary, based on: Chongkan Songben shisan jing zhushu: fu jiaokanji, Taibei, Yiwen yinshuguan (or other identical reprints)
The page numbers given in the database refer in both cases to the editions stated.

5.2 Standard search with simple search statement

a) select from the menu 2) 自由檢索
b) select from the appearing menu 1) 檢索
c) select 2) 直按輸入
d) enter the search phrase with the input method provided by your system; do not forget to switch back to English input afterwards
e) press the "Enter" key, select in the appearing frame 輸 入, enter the search range with a path statement (compare above); for
paths go back to the menu choice 1) 目錄與正文 in the very first screen.
Examples:

In all databases:
/   will search the whole database

In the database of the 25 histories:
/2  will search only the Shiji (Shiji is /2, the Qing Shigao is /26)
/2-4  will search Shiji, Hanshu and Houhanshu

In the database of the 13 classics:
/1  will search all the 13 classics with commentary
/1.2  will search the Shangshu with commentary
/2  will search an interpunctuated version of the whole 13 classics without commentary
and so forth.

f) press the "Enter" key, wait for the search to be executed
g) select B) 查閱 from the menu
h) press "g" to go to the first search result, press "g" again to get to the next search result, "G" to move back to the previous one,
"Enter" to get to the next section of the continuous text
i) press "t" to get back.

5.3 Standard search with phrase exclusion in the search statement

When entering a search statement, any phrase consisting of a character enclosed in braces { } and the character(s) immediately preceding or following the braces will be disregarded.
Example: To get mostly passages which contain 風 in its literal sense "wind" but not in the derived senses "custom", "style" and so on, type:
{ 國 扶 遺 鄕 承 歌 八 移 } 風 { 伯 后 師 土 俗 教 刺 諫 喻 }.

5.4 Standard search with complex search statement (Boolean operators)

Search statements can not only consist of any number of Chinese characters, but also of Boolean operators to determine the logical relationship between the characters, round brackets to determine the priority of the Boolean operators: Boolean operators

& AND
! AND NOT
| OR

In the absence of brackets, & as well as ! have priority before |. Round brackets ( ) can be used to change the priority of the operators. The operators are valid within one text section, i.e. the lowest hierarchical unit in the directory structure. It does not matter if there are spaces before and after the operators or not.
Example: To research disaster relief policies, input
(減免 | 除) & (田租 | 口賦).
This will give all text sections that contain any of the two phrases 減免 and 除 combined with any of the two phrases 田租 and 口賦.

5.5 Inclusion/exclusion of specific text levels (e. g. commentary)

a) enter the retrieval system, choose a database (see 5.1 above); if you are already in the retrieval system, go back to the very first screen (usually by pressing repeatedly "t")
b) select from the menu 1) 目由與正文
c) press "c" to choose content type (e.g. headings, commentary, philological notes) only the content type chosen will be displayed and searched
- in the frame that appears, use the following commands:
"m" (mark) to select and deselect, "q" or "t" (quit, terminate) to save and exit
- available content type:
* 全部


券標


朝代

序言
段落含詿


詿釋群
詿釋
勘記
不詳
The de-selection of any upper-level content type will also exclude all content types below, e. g. de-selecting 段落含詿 will also exclude 段, 詿釋群 and 詿釋 from display and search, even if not explicitly indicated. De-selecting 券 will exclude the entire content of that 券 and so on. Use the d command (e. g. 7d) to get an idea about the hierarchical relationship between different content types.
d) press "t" to save and exit
e) press "t" again to return to the very first screen
f) choose from the menu 2) 自由檢索 and execute any type of search (see 5.1, 5.2, 5.3 and 5.5); the selection remains valid for all further searches and navigation until you explicitly select new content type or choose a new database.

5.6 Index searches

This type of search leads to a different result handling which greatly simplifies the evaluation of large result sets and is ideal for linguistic/terminological questions:

a) From the very first screen, select 2) 自由檢索
b) Select F) 式
c) In the frame that appears, select 引得檢索, press "Enter"
d) Select 1) 檢索
e) Select 2) 直按輸入, as for a simple search
f) Enter the search statement and select the search range, as above (5.1 d-f)
g) After the search has been executed, select B) 查閱 from the menu
h) For sorting the result list, first press "t" and then select from the menu S) 排序
i) In the frame that appears, select in which priority you want to have your search result sorted; then press "Enter"
j) If you want to examine the whole passage, press "g", as in a simple search (5.1 h)

To perform a standard search, choose in step (c) 一船檢索 instead of 引得檢索.

NOTES:

  • The index search can handle a maximum of 5000 search results.
  • For an index search, only one phrase (consisting of one or more characters) can be inputted; if you input a search statement that contains Boolean operators, only the first phrase will be taken into account, all other phrases will be disregarded. If you want to a Boolean search with an index search, first perform a standard search with the complex search statement, switch back to index search, re-input only the most important phrase of your previous (complex) search statement and then choose your previous result set as search range by selecting 重疊檢索 in the frame where you define the search range (compare 5.1 e; see also below).

5.7 Search on previous result set and combination of searches

To perform a search on your previous result set, select selecting 重疊檢索 in the frame where you define the search range (compare 5.1 e; see also above). This can be done several times. After the first time, you will be asked which result set you want to use. Choose be inputting numbers.
Performing in this way overlapping searches is helpful for speeding up the search, refining your search without starting all over again, avoiding overly complex search statements, combining index searches with complex search statements (see 5.5) etc.


Top | Introduction | Telnet Version | Special Characters | Navigating the Database | Searching the Database | Further Information

6. Further information

Further information can be obtained from our homepage http://www.sino.uni-heidelberg.de/.
You are welcome to contact us at library.uni-heidelberg.de.


Last update: Aug 13, 2003 (MA)