KR19990039091A - How to define dictionary and text corpus structure in text dictionary management system (TDMS) - Google Patents

Abstract

The present invention relates to a dictionary and text corpus structure definition method in a text dictionary management system (TDMS), and in detail, the regular structure of the dictionary so that it can describe various types of electronic dictionary logical structure as effectively as possible using SDML (Constant format part) is defined in SGML DTD form and must follow the format. The part related to the rest of contents (parts that can be changed) can be freely defined by SGML. By defining the dictionary and text corpus structure for text corpus management by allowing text corpus to be stored as simple text, it can be used as a standard for Korean documents as well as foreign documents. To manage, transform and transform it into new information One dictionary and text corpus structure definition method.

Description

How to define dictionary and text corpus structure in text dictionary management system

The present invention relates to a method for defining a dictionary and text corpus structure in a text dictionary management system (TDMS), and in particular, defines a dictionary and text corpus structure using a standard dictionary markup language (SDML). In order to be able to describe various forms of electronic dictionary logic structure as effectively as possible, the regular structure of the dictionary should be fixed form, and only the parts related to the rest of the contents can be freely defined, and the text corpus can be stored as simple text. The present invention relates to a dictionary and text corpus structure definition method in a text dictionary management system (TDMS) capable of text corpus management.

In order to effectively transform and manipulate a large amount of document information recorded by a computer, the document format must be unified. This document format unification can be referred to as Standard Generalized Markup Language (SGML) as can be seen in the following reference. The trend is unifying.

<Reference>

1. Industry Promotion Agency (1993), "Document description language SGML," KSC 5913-1993, Korean Standards Association.

Bryan, M. (1988), SGML: An Author's Guide to the Standard Generalized Markup Language, Addison-Wesley.

Goldfarb, Charles F. (1990), "The SGML Handbook," Claredon Press, Oxford.

4. Travis, B. E. and Waldt, D. C. (1995), "The SGML Implementation Guide," Springer-Verlag Berlon Heidelberg, Germany, pp 3-20.

SGML is a meta language that defines the syntax of the markup language. You can use it to define various document structures. In other words, SGML defines a grammar structure that can define a syntax structure, and the syntax structure is defined using the DTD (Document Type Definition).

Figure 1 shows the definition of the element declaration portion of the SGML.

Where mdo is used to indicate the start of a markup definition, where "<!" Is used. "Element" means to define an element element of SGML as a keyword. ps means blank, ps + means one or more blanks ps * means none or more than one blank.

element type is the name of the element you want to define. omitted tag minimization indicates whether the start and end tags can be omitted from the currently defined element type. (declared content | content model) means either the declared content being used inside SGML, or one of the content models that define the structure of the document. "|" Means that one of the two is selected. Finally mdc is used to mark the end of the markup definition.

According to these SGML description rules, various document formats can be defined as DTDs. For example, the DTD of SGML, such as analysis method 2a, is intended to mark up the example document as analysis method 2b.

In other words, SGML is a grammar of the language (DTD) that can define the structure of the document, DTD is the structure of the document according to the SGML rules, mark the general document using tags defined in the DTD according to the structure One is the markup document. In general, documents marked up according to the SGML DTD are also called SGML documents.

For documents in various fields written in text, the TEI (Text Encoding Initiative) proposes several standard forms, such as those referenced below.

<Reference>

5. Ide, Nancy and Sperberg-McQueen, C. M. (1995), "The TEI: History, Goals, and Future," in Ide, Nancy and Veronis, Jean [20], pp 5-15.

6. Ide, Nancy and Veronis, Jean (1995), editors, "Text Encoding Initiative Background and Context," Kluwer Academic Publishers, Dordrect / Boston / London.

7. Sperberg-McQueen, C. M. and Burnard, Lou (1994), "Guidelines for Electronic Text Encoding and Interchange (TEI P3)," Vol I and Vol II, ACH, ACL, ALLC.

8. Sperberg-McQueen, C. M. and Burnard, Lou (1995), "The Design of the TEI Encoding Scheme," in Ide, Nancy and Veronis, Jean [20], pp 17-39.

For example, Analysis Method 3 shows some of the dictionary-related DTDs defined by the TEI, which can be in any order, such as home, sense, form, gramGrp, usg, def, etym, eg, after the entry tag. As a structure that can be used over and over again.

However, the standards of these references are based only on the type of document, especially Western documents, and the definition of the information required to manage this document is based on the methods described in the following references.

<Reference>

9. ISO / IEC (1994), "Information technology-Text and office system-Document Style Semantics and Specification Language (DSSSL) -Draft", ISO / IEC DIS 10179.2.).

It is problematic to accept and use such foreign document-oriented standards as standards for Korean documents, and there are limitations in managing and transforming documents in the form of dictionaries that contain aggregated information. In the case of TEI, we can interpret home as a synonym, sense as a meaning, and use it as a tag, but we need to define it correctly to fit our dictionary. In particular, tags indicating "academic field", "di dialect", etc., which appear in the Korean dictionary, are missing, and the divisions of Chinese characters, foreign languages, Korean, etc., are missing from "the etymology."

In other words, TEI defines the structure of a dictionary, which is very generally applicable, but there is a problem in applying the Korean dictionary because there is no consideration of the Korean dictionary as described above.

Therefore, it is justified that a standard notation format required for newly constructing, managing, and transforming a dictionary document into new information has been emphasized.

In order to solve this problem, Kang Bum-mo proposed a new structure suitable for the Hangul dictionary, and if this part of the structure is displayed hierarchically, it is the same as the attached analysis method 4 (Reference: Kang, Beam-mo (1996), "Using the TEI Scheme in compiling a Korean Dictionary ", ALLC-ACH '96, Bergen, Norway.)

However, since the gangbeom method, which has the same structure as TEI or analysis method 4, considers too many cases, it is too computationally complex to be processed in the actual text dictionary management system, and the parsed module and the parsed result. There are a lot of difficulties with using.

To explain this in more detail, in the analysis method 4, the hom in the next step of the entry is recursive under the hom, and thus, the step may theoretically be repeated infinitely. While this process can technically reduce the complexity of the computation by specifying the iteration limit—the processing overhead of parsing each step and processing its contents—but still requires a large amount of computation. do. In addition, if the TEI document is defined in accordance with checking whether the TEI document is defined according to the grammar rules (DTD) defined in the TEI, the storage form of the parsing module that converts and stores each content into a form that can be easily processed by a computer is usually a tree. As it is expressed as a structure, when iterative definition is made as described above, it is necessary to consider the possibility of being used repeatedly indefinitely, which causes a problem of creating an infinite tree structure. In order to avoid this, since the lower node of the tree must repoint the upper node, a cycle is actually formed, and it becomes a form other than the tree, which requires another complicated process.

In addition, there is an analysis method that can be understood as a structure different from the structure of TEI or Kang Beom Model, that is, the structure shown in Analysis Method 5, as shown in the following reference. It cannot represent a wide variety of forms (intermediates rather than various forms that can lead to complex implementations, such as iterations, as shown in the TEI or Kang Bum Mo approach). Therefore, there is a problem in accepting this as a general dictionary DTD.

<Reference>

10. Choi Byung-jin, Lee Jae-sung, Lee Un-jae and Choi Ki-sun (1996), "Logical Structure of General Dictionary for Standardization," 8th Hangul and Korean Information Processing Conference.

11. Byung-Jin Choi, Jae-Sung Lee, Un-Jae Lee, Ki-Sun Choi (1996), "Logical Structure of Dictionary Items for Machine-Ready Dictionaries," Cognitive Science 7-2.

In addition, the existing dictionary structure definition method has no definition of how the management system will display the display of each tag. This makes it possible to create more diverse management systems, but there are problems in using them as standard management systems.

An object of the present invention for solving the above problems is available as a standard of Hangul documents as well as foreign documents, text dictionary management system suitable for constructing a new dictionary, managing and transforming it into new information (TDMS). It provides a method for defining dictionary and text corpus structure in.

1 is a diagram illustrating a definition of an element declaration portion of an SGML.

2A is an exemplary diagram showing an SGML DTD

FIG. 2B is an exemplary diagram of a document marked up with the SGML DTD of FIG. 2A

3 is a diagram illustrating some of the dictionary-related DTDs defined in TEI

4 is a diagram illustrating a part of the Hangul dictionary structure proposed by Kang Bum-mo

5 is a diagram showing a part of the Korean dictionary structure

6 is an exemplary diagram of a dictionary structure expressed in SDML according to the present invention.

7 is an exemplary view illustrating the redefinition of the substructure of <body> of FIG. 3.

8 is an exemplary diagram of a text corpus structure expressed in SDML according to the present invention.

9 is a diagram illustrating a method of defining an SGML content model.

Features of the present invention for achieving the above object is to form a fixed structure of the dictionary in order to describe the electronic dictionary logical structure of various forms as effectively as possible, and the parts related to the rest of the contents can be freely defined, By defining the dictionary and text corpus structures using SDML, which allows text corpus to be stored as simple text and manages text corpus.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

First, the SDML used for the definition of the structure of the dictionary and the text corpus of the present invention is defined by the SGML DTD form a certain format portion and must follow the format freely by using the method to define the format can be changed in SGML However, it can be defined by the SDML restriction rules.

Such SDML includes a header area that contains contents such as the name and version of the dictionary, a definition area that defines the values each attribute can use, and the contents of the actual dictionary. Each entry is composed of a repeated entry group area or a text corpus file group area in which the contents of a text corpus file are repeated.

6 and Analysis method 8 show the dictionary and text corpus structure hierarchically using the SDML as described above. The standard dictionary term is the top layer, and the top layer is the heading of the standard dictionary. Includes in its sublayer a term containing information of a term, a term defining tag attributes that can be used in a user-defined tag of a standard dictionary, and a term of a dictionary entry or a set of text corpus files in its sublayer. Includes subheading terms and descriptive terms attached to dictionary headings.

The terms or names shown in the following Analysis Methods 6 to 8 will be briefly described with reference to Table 1 below.

tag Function & role <sd> Standard dictionary term <sdHeader> The heading of a standard dictionary that contains information such as the name and version of the dictionary. <sdFront> Terms that define tag attributes that can be used in custom tags in the standard dictionary. <group> A set of dictionary entries or files <entry> Dictionary entry terms <wname> Heading terms in dictionary entries <body> Terms that include descriptions in dictionary headings <tdmsfile> Terms containing the contents of the file

In general, when using a term or name as a tag in the definition of a document structure, it is distinguished from other data by enclosing it with <>, which is enclosed with <> to indicate the beginning of the item, and enclosed with </> indicates the end of the item. Indicates.

In this case, the terms and names as described above are actually defined in the DTD as shown in FIG. 7, and may be defined differently according to the intended purpose.

However, in the present invention, the following tags are defined as special and are used separately from general tags (tags that can be defined by the user).

That is, <sd> represents the entire standard dictionary, and indicates that the elements include <sdHeader>, <sdFront>, and <group>. In addition, <entry> represents a subterm of the <group> term and includes <wname> and <body> as subterms (see FIG. 6).

Similarly, as shown in FIG. 8, the subterm of the <group> term may include the <tdmsfile> term instead of the <entry>, meaning that the term contains the contents of a text corpus file, not a dictionary entry. .

Therefore, the structure of the dictionary and text corpus according to the present invention will be described based on the matters shown in Table 1 as follows.

The <sdHeader> representing the header area contains contents such as the name and version of the dictionary. In addition, <sdFront>, which defines the definition area, defines the value that each property can have in the dictionary, and <group>, which represents the collection area, contains the contents of the actual dictionary, and each <entry> that contains the headword. Or an iteration of a <tdmsfile> containing the contents of a text corpus file.

In the present invention, only the parts related to the contents of the dictionary structure defined in FIG. 6 can be freely defined. The freely definable [standard dictionary element] part is an element included in the <body> and is a simple one as a default value. It can be composed of fields, or it can have a complex tree structure.

To do this, the user can define the structure of the necessary dictionary and change the structure of the <body>.

At this time, the <body> is initially defined as follows.

<! ELEMENT body ........ (# PCDATA) *>

Therefore, when changing to a new structure, only the above elements are overwritten.

For example, as shown in FIG. 7, the <body> structure may be changed to a new structure including <part of speech> and <band word>. The redefinition of such a structure requires each application program. You can use the structure and tag of.

At this time, the definition of the new structure is generally defined in the same way as the format of SGML, but there are some rules that should be used limitedly in SDML. The general rules are as follows.

Condition 1: SDML basically expresses the structure of a document in a tree form using the SGML content model definition method of FIG. 9. The basic elements that can be used are sequence (,), alternation (|), repetition (*), one or more repetition (+), and optional (?). At this time, the free sequence (&) is excluded because it is meaningless in determining the dictionary form.

Condition 2: Do not allow the recursive structure of tags that can appear in the presentation of the document structure.

Condition 3: The name of the tag is used as the field name in the editor that edits the document defined in SDML (SDE: Standard Dictionary Editor).

Condition 4: When using the display information used in the editor, display it by including it in the tag using the attribute name and attribute value specified in the TDMS. Unless otherwise specified, display information is displayed in the editor in the default display manner.

Here, the display information is information used to display each field on the screen, and indicates the font size, font shape, size and type of the input window or display window, and the property is a property or tag that can be used internally in the dictionary editor. It is used to define things that are a bit inappropriate for expression.

For example, if the <synonyms> tag is displayed on the screen in the dictionary editor, the contents of the <synonyms> tag can be underlined and the headings of the synonyms can be written as attributes of the <synonyms> tag.

In addition, the default display method is to use a single line input window by default when inserting contents in a tag of an element type defined as #PCDATA. It can also vary from implementation to implementation, but it defaults to the default value initially assigned by the editor.

Condition 5: If a tag is already defined, use that tag name as much as possible to ensure compatibility.

So far, we have looked at some rules that should be limitedly used in SDML. The condition 5 is actually for standardization of tags.

Hereinafter, the management of the text corpus, in the present invention, in the definition of the dictionary structure expressed in SDML to manage the text corpus, <entry> is replaced with <tdmsfile> so that the text corpus can be stored as simple text in this tag. The relevant information is recorded in the header to distinguish the text corpus from the dictionary. The structure of the text corpus expressed in SDML is as shown in FIG.

In the SDML used for the dictionary and text corpus structure definition of the present invention as described above, most of the structure expression format of SGML is used as it is so that it can describe various types of electronic dictionary logic structures as effectively as possible. That is, it uses the definition of ELEMENT structure of SGML as it is and restricts the use of unnecessary functions. In addition, in order to keep the definition of the standard tag as uncertain as possible, and to manage the electronic dictionary as efficiently as possible, the regular structure of the dictionary should be fixed and only the parts related to the rest of the contents can be freely defined. According to this principle, management information can be included in the current SGML tag format and processed. In other words, the display information is included in the attribute so that the structure can be freely defined using a tag.

The dictionary and text corpus definition method of the present invention as described above is effective in newly constructing, managing, modifying, and transforming new documents into dictionary information by limiting some functions and considering information used in the management system. Can be used as a language for standard dictionaries.

Claims (4)

In defining the dictionary and text corpus structure in the text dictionary management system, the regular structure (constant format part) of the dictionary is defined in the form of SGML DTD to be in a fixed form, and the parts related to the rest (the format may be changed). Part) can be freely defined using the method defined in SGML, and the dictionary and text structure in the text dictionary management system characterized by using SDML which can manage the text corpus by storing the text corpus as simple text. Definition way. The method of claim 1, The SDML includes a header area including contents of a dictionary name and a version, a definition area defining values that each attribute to use in the dictionary, and the contents of the actual dictionary, each containing a headword. It consists of a repeated entry group area of entries or a text corpus file group area in which the contents of a text corpus file are repeated. The standard dictionary term is a top layer, and the top layer is a header of a standard dictionary, such as the name or version of a dictionary. Include in its sublayer a term containing information, a term defining tag attributes that can be used in a user-defined tag of a standard dictionary, and a set term of a dictionary entry or text corpus file, the dictionary entry set term being a subordinate thereof. Defining a hierarchy to contain a term containing a term and a description of the dictionary term How to pre-define and structure of the text in a text dictionary management system as claimed. The method according to claim 2, wherein the user can redefine the structure of the term including the description attached to the dictionary headword. 4. The method according to claim 3, wherein the redefinition of the structure of the term including the description attached to the dictionary heading is performed by using the ELEMENT structure definition of SGML.