SD - SGML

We have the evolution of:

Markup language, to
Metalanguange / generalized markup, to
SGML (metalanguage) and
Application of SGML using
Document Type Definitions (DTDs).

SGML - Standards

SGML = ISO 8879, Standard Generalized Markup Language
Related Standards
- ISO 8613, ODA, Office Document Architecture: formating, imaging, interchange; represented in SGML or binary --- includes layout
- ISO 10179, DSSSL, Document Style Semantics & Specifications

SGML - Explanation

Is a
- tagging language, DB language
- extensible doc. descrip. language
- metalanguage to define doc. types
Supports
- logical structure, hierarchies
- file linking & addressing
- multimedia and hypertext

Documents, DTDs, Procedures

SGML-encoded documents can be:

checked for syntactic compliance with a DTD
printed in accordance with DSSL specifications
displayed on a screen
indexed for context-specific searching
translated into another representation

SGML - Markup

SGML allows 4 types of markup

Descriptive: tags
Referential: entity references
Metamarkup: markup declarations
Processing instructions: LINK, CONCUR

The most important type is Descriptive Markup.

Elements

GI (generic identifier) - found in start and end tags
id= idref; attribute/value pairs
content

Top element = doc.
Elements have content model (i.e., a grammar production)

Example

<!ELEMENT textbook - - (front, body, rear) > <! ELEMENT fig - - (figbody, figcap?)> <! ELEMENT figbody - O (artwork | (p | ol | ul)+)> <! ELEMENT artwork - O EMPTY> <! ELEMENT figcap - O (#PCDATA)> <!ATTLIST fig id ID #IMPLIED> <!ATTLIST artwork depth CDATA #REQUIRED> <fig id=ohcofig> <figbody> <artwork depth=24p> <figcap>OHCO Model Diagram </fig>

DTDs

Define a class of documents.
Specialize SGML for documents in a class.
Use an attribute grammar.
Use a bracketted grammar.
Hierarchy support means nested content.
Ordering means OHCO is fully supported.
Developing a DTD: DTDspin, modular approach, DTD for theses, ACM's eclectic approach (but making sure formatters exist).

Our Uses of SGML

CODER: dictionaries, INCARD (project with cardiology data)
Project Envision

Bibliographic records
Transactions articles

Tech reports (WATERS): for biblographic records instead of refer or BibTeX
Dissertations and theses
TREC, TULIP
HTML and HTML+

Advantages of SGML

Maintainability
- Allows multiple use: Paper - screen; thesis - article - book
- Immune to new WP s/w or new versions
Portability
- Author to online use to publisher
- Electronic Manuscript Project (of AAP)
  - Guidelines for articles, books
  - Elements for mathematics, tables
- Text Encoding Initiative (TEI)
  - Proposed and run by: Humanists, Linguists, Literary Community
  - Guidelines for Electronic Text Encoding and Interchange (P3): dictionaries, drama, poetry, sculpture descriptions, simultaneous translations, ...
Benefits
- Sharing, collaborating
- Avoid rekeying, proofing, fixing galleys: initially, and on new editions
- Derive bibliographies & citations
- Immediately publish too in online databases and on CD-ROM
Less Cognitive Demand, Better Support of Content Orientation
- Element recognition only -> derive markup selection & performance
- Focus on structure
Special Assistance / Processing
- Multiple views (recall LectorMotif)
- Outlining / structure editing (e.g., Author/Editor)
- Panorama: SGML (presentation) on the Web

Doc. Modeling w. SGML

Characterize structure using DTD

Depth / thoroughness of tagging
Bibliographic entries, names
Math, tables

PAT model, representation, retrieval
PAT and LECTOR specification files
OHCO, grammar models --- especially if one avoids SGML customization
- Minimization, entity references
- CONCUR, processing markup
- Extensive use of attributes
Attach semantics / weighting / evidence effects -- to context