Xerox sets its sights on distributed authoring

May 5, 1998

Liora Alschuler

The Seybold Report on Internet Publishing
Vol. 2, No. 9
Liora Alschuler
May, 1998

Xerox's Raven is a prototype of an XML editor developed as a research project within one of Xerox's technical publications departments. The design objective is to target the low end of the editing spectrum with a tool that is simple to install and maintain, supports forms-based data entry, and is effective in a distributed authoring environment. In targeting writing rather than publishing, we feel that Xerox is staking out a valuable piece of property. How many writing licenses support one publishing license, after all?

The Raven software, debuted at XML '98's Developer's Day, is not much more than a proof of concept. What we were able to see were a clean user interface and innovative problem-solving techniques applied to areas that have gone begging for too long. For example, one of the dirty little secrets of most (though not all) SGML implementations is that tables, which clearly are relational data, are marked up according to their appearance (e.g., this lays out in a <column>; this is a <row>). Appearance-based markup neutralizes the value of the information for search and retrieval and other purposes.

click to load illustration

Raven supports forms-based data entry, where each element is tagged according to its meaning (e.g., this is a <price>; this is a <quantity>). The semantically tagged information can have behavior associated with it. That behavior includes tablular functions, such as performing calculations-as well as lining up rows and columns.

Running in Java.

Another innovative feature of Raven is that it is being written as a Java applet. While several products support distributed Web authoring, Raven would be the first XML editor to do so. Running as a Java applet has several apparent advantages, most notably in the areas of administration (only one copy of the program to maintain for an entire workgroup) and access control (check-in/check-out controlled on the server).

click to load illustration

Raven can also provide a direct link from the document to workflow or post-editing processes. In fact, the use of Java means that the document itself can become an application interface, similar to the way macros work in word processors, but with the full power of the Java programming language in a validating, structured editor.

Xerox is targeting applications such as collaboratively edited status reports, multimedia scripts, and forms. Currently, the processing overhead of one Java class per element type puts a ceiling on the effective length of a document. Xerox is considering using XSL to transfer formatting to declarative markup, making it more accessible to the end user.

The team that nurtured Raven thus far is in discussion with groups inside and outside of Xerox to determine the need for such an editor and, given an identifiable need, to locate the resources required to bring it through development to market. (For more information, contact Scott Parnell at