Chrystal’s Canterbury Moves Component-Based Information Management out of the SGML Realm
Chrystal Software is a Xerox New Enterprise company-a part of Xerox that supports the development of new technologies. Documentum, one of the premier document management systems, also began as a Xerox New Enterprise. We first encountered Chrystal’s products in their earlier manifestation at X-Soft. The product that later became Chrystal’s Astoria was developed at Xerox PARC and marketed by X-Soft as InContext. Astoria, Chrystal’s initial SGML-based product, has been marketed to customers in the aerospace, automotive, and telecommunications industries. When Chrystal was formed, Astoria was firmly settled in the SGML world. If one didn’t want to pursue SGML, Astoria held little interest.
Now, the component-management system at the heart of Chrystal’s design has opened new levels of exciting possibility. With the introduction of Canterbury, we no longer have to go the SGML route to achieve our goals. To quote Chrystal’s product literature:
Studies show that at least 30 percent of the content created by technical publishing groups could be reused in other documents. Then why are authors `reinventing the wheel’ again and again? The problem is authors can’t find the information, which may be buried inside any number of documents. And the process of opening up these documents, and cutting and pasting information into new documents is so tedious and labor-intensive that there may be very little actual time savings over writing from scratch. (Canterbury Reviewer’s Guide, pg. 3).
Canterbury supports FrameMaker 5.5.6, eliminating the need for SGML coding. Instead, Canterbury deconstructs FrameMaker documents using the standard FrameMaker tags to identify and export components in XML to an object-oriented database.
Canterbury promises to take existing manuals and have them automatically separated into components that can be reorganized into manuals, Web content, online help, and CD-ROMs-all from the same content. Guided by tags assigned in FrameMaker and additional meta-data added to the database, Canterbury breaks the documents into components that are ready to be reused.
Using these capabilities, we can create documents as we would in FrameMaker, using conditional text and variables to assign unique components based on platforms, customization, user requirements, media, and so on. We can export the documents to the database where they are broken down into information at the paragraph level. Then, we view the FrameMaker files as virtual documents containing images of the information in the database. By searching for and selecting components that already exist, other writers can create new documents using existing components in addition to new information objects they add. We can lock components so that only designated people can revise the content. We can reuse the same components in multiple documents; if they are changed in one place, they will change everywhere they are reused. We can even reuse the same components within a single document simply by referring to the unique object in the database.
Here are some of Canterbury’s features as they relate to the needs of technical publications organizations:
Granularity to the paragraph level-Canterbury automatically breaks FrameMaker documents down to the paragraph or file level.
Updates-Canterbury updates all documents that have reused a particular component when that component is revised. Although a component may appear to have been copied into a document, what has actually been inserted is a pointer to the component. From the Repository Manager, you can view a list of all documents that include a particular component. If you have 500 documents using the same component, you need only change the component once and every instance of that component is immediately updated in all 500 documents.
FrameMaker-Canterbury works with FrameMaker 5.5.6 files. Writers need not learn a special tagging language. Canterbury breaks documents into components based on tags already assigned in FrameMaker.
FrameMaker variables-If you set up a variable with FrameMaker, that variable will extend to any document that reuses a component related to the variable. Canterbury also maintains all your FrameMaker cross references.
Searches-Canterbury’s search engine allows writers to search for a component by its text or by other criteria based on its function or location in a document, or by any of the meta-data that has been added to identify the component in the database.
Attributes (meta-data) can be assigned to components to aid searching-Typical attributes might be subject matter (troubleshooting, maintenance, etc.), status information (in review, final draft, etc.), and user type (novice, expert, etc.).
Security-Supervisors and project managers can assign access privileges to control who has authority to change a component.
Check in/Check out-Only one individual can check out and work on a component at a time.
Change history-Canterbury allows writers to enter comments about the changes they are making to a component.
Notification-Canterbury notifies the designated author of changes made to a component.
Version Control-With component-level version control, you are able to see which part of the file was changed and why (through notes attached to that component).
Multiple format storage-Like most document management systems, Canterbury will store images, video clips, spreadsheets, presentations, and other file types.
Chrystal offers an extension product called Lingua to manage multiple language versions of your documents and reduce translation costs. All the language versions are stored in the database with links to the base language components. If a writer changes a paragraph in the original language, the system identifies those paragraphs in the translated versions that will have to be retranslated. Such a system should work particularly well with translation memory systems, such as Trados.
Chrystal provides a web tool that makes versions of the database components available through intranets, extranets, or the Internet. Users can search for particular components and display them in a browser. Functionality of this type is particularly useful to support help desks. The Web capability can also be used for reviews. Reviewers can attach comments to a component directly. The comments are identified by date, time, name, and other required information.
Most of the products that are well-known in the information management market today are document management systems such as Documentum and PCDocs. These systems allow us to manage at the file level and offer sophisticated tools for version control, security, check in/check out, and change history. They also offer workflow capabilities, which Canterbury does not. Using workflow setups, you can automate the process of sending documents out for editing and review. You can, of course, add a standard workflow tool to Canterbury, if needed.
Products such as Hynet offer publishing options superior to those that Canterbury appears to offer at this time. Using Hynet’s Directive (reviewed in Vol. 1., No. 1 of Best Practices), you can output derivative documents from the database to a variety of media such as HTML, print, XML, and Directive’s electronic book. With Canterbury, you use the publishing capabilities in FrameMaker but once your document is outside the database, it is no longer handled by Canterbury’s control mechanisms.