AIS does SGML-to-HTML conversion

January 10, 1998

Seybold Report on Internet Publishing
Vol 2, No 5
January, 1998

HTML Package, an add-on to Balise from the French company AIS Software, converts large SGML documents to Web-ready HTML. The two major components of the add-on are a style sheet editor and a batch conversion utility. The style sheet editor creates XSL-like style sheets using the Balise programming language instead of ECMAscript. Style specifications can be constants or expressions. The editor will be brought into XSL compliance as the new standard emerges. The interactive editor, however, is designed for SGML fragmentation and conversion, thus it addresses issues that lie beyond the scope of a generic design tool.

Microsoft IE is executed as an ActiveX control within a Windows application to provide a real-time preview of the rendered HTML. The resulting files, however, work within any level 3 or above Web browser.

Site navigation is inserted into the HTML pages automatically (see photo above).

Creating HTML from SGML with HTML Package is a two-step process. First, one takes a bird’s-eye view of the source document looking at the size of each element instance and the element tree. HTML Package provides a hierarchical view of the source indicating the size of each element instance. Then, using the WYSIWYG style editor, the user decides where to divide the HTML rendering into pages.

Users can generate text in HTML and sort and reorder the fragments, using what François Chahuneau, president of AIS Software, calls "node-based re-ordering by tree-based access." So, for example, a user can gather coded bibliographic citations interspersed within an SGML source document, sort them, link them to their references and present them on a single HTML page. Since chunking and reordering is done on the fly from the style sheet, each SGML source can have multiple renditions. In the HTML Package way of thinking, HTML navigation aids are a manifestation of style, as are decisions on the size of the page delivered.

The style sheet editor supports:

  • Manual and rule-based fragmentation.
  • Inter- and intra-document links.
  • Automatic generation of navigational links between fragments.
  • Automatic generation of multiple tables of contents.
  • SGML CALS tables.
  • HTML frames.
  • All Balise programming functions.

For preview purposes, the user can generate HTML on-the-fly from the source plus style sheet. For delivery, AIS recommends precompiling the HTML. The intent is to provide interactive design for static generation of HTML pages.

HTML Package is available now and is priced at $2,500. It requires a Balise Developer’s License (see below). The style sheet editor runs under Windows 95 and NT and requires Microsoft Internet Explorer 3.0 or above. The batch conversion engine runs on any platform that supports Balise.

Dual Prism from AIS. AIS Software also announced and demonstrated beta software of a second Web delivery product called Dual Prism. Dual Prism will build on the Balise/HTML Package concept with these important extensions: It is designed for real-time rendering from disparate data sources with structured and full-text queries. The image of two prisms depicts data from relational tables, SGML markup, RTF documents and HTML converging into the Balise engine and multiple style-sheet-based outputs diverging on the other side. RTF integration will be based on the RTF Scanner (see Utilities and Programming Tools further on) and database integration on an ODBC interface. In response to a query, the interface will pass the relational data to Balise, which will generate HTML tables. The package will include load balancing and a multithreaded server.

Dual Prism is designed as a real-time delivery engine and will not attempt to replicate or replace standard management functions such as version control and check-in/check-out. Application of styles will be based on the same principles as HTML Package, but operating on memory- or disk-based storage without size constraint.

AIS had three major announcements for products either being shipped or announced for shipment within the next six months. It appears to be targeting markets dominated by the most successful North American SGML companies: OmniMark’s conversion and programming market and Inso’s data conversion and Web publishing market. While it has established a reseller network that includes three offices in the U.S. and Canada, AIS has not "Americanized" its product marketing literature, which still sets it apart from the U.S. mainstream. It will take a major investment to bring RTF Scanner up to where it can compete off the shelf with DynaTag, if that is the intent.

Designed as a turnkey solution for nonprogrammers, Dual Prism is an ambitious project. While the Balise programming language is no doubt capable of the task, the software will be judged on its ability to condense the flexibility and power of a programming language into an end-user-level GUI.

AIS anticipates pricing for Dual Prism will be in the range of $10,000–20,000 for a minimum configuration scalable to hundreds of users. It is scheduled for release by the end of June, 1998. See also coverage of Balise Release 4.0.