Features summary

• Includes nsgmls
  • Compatible with sgmls
  • Also generates RAST (ISO/IEC 13673)
• Provides access to all information about SGML document
  • Access to DTD and SGML declaration as well as document instance
  • Access to markup as well as abstract document
  • Sufficient to recreate character-for-character identical copy of any SGML document
• Supports almost all optional SGML features
  • Arbitrary concrete syntaxes
  • SHORTTAG, OMITTAG, RANK
  • SUBDOC
  • LINK (SIMPLE, IMPLICIT and EXPLICIT)
  • Only DATATAG and CONCUR not supported
• Sophisticated entity manager
  • Supports ISO/IEC 10744 Formal System Identifiers
  • Supports SGML Open catalogs
  • Supports WWW
  • Can be used independently of parser
• Supports multi-byte character sets
  • Parser can use 16-bit characters internally
  • 16-bit characters can be used in tag names and other markup
  • Supports ISO/IEC 10646 (Unicode) using both UCS-2 and UTF-8
  • Supports Japanese character sets (Shift-JIS, EUC)
• Object-oriented
• Written in C++ from scratch
  • Not a modified version of a parser originally written in C
  • Reentrant
  • Sophisticated architecture
• Fast
  • Up to twice as fast as sgmls on large documents
• Portable
  • All major Unix variants
  • MS-DOS
  • Win32: Windows 95/Windows NT
  • OS/2
• Production quality
  • Version 1.0 recently released, after a year of test releases
  • Tested using several SGML test suites
  • Already used in several new commercial products
  • Written by James Clark, previously responsible for turning arcsgml into sgmls
• Free
  • Includes source code
  • No restrictions on commercial use
• Disadvantages
  • Programmer-level documentation only for generic API and not for native API.