Inso releases DynaBase 2
January 20, 1998
The Seybold Report on Internet Publishing
Vol. 1, No. 9
Inso's DynaBase, which has been in limited customer testing for what seems like forever, now looks ready for commercial deployment.
DynaBase is a client-server system for collaboratively authoring Web sites that have a mix of static and dynamically generated content. In addition to providing the typical features of document management (check-in, check-out, version control, configuration management), it provides a framework for applying high-end document management to Web publishing. That means secure, remote collaboration over http.
Typical installations run in the $45,000 to $50,000 range, with at least 1,000 pages and a team of a minimum of three people.
Data server. The repository, based on the Object Design object-oriented database, stores all of the elements of the site-text, graphics, multimedia elements, templates, scripts-providing version control and secure access to all of these elements. Like other document-management systems, DynaBase assigns attributes to all of the objects. In contrast with some of the new low-end solutions (such as DocuShare, covered elsewhere in this issue), these attributes are fully customizable by the customer.
DynaBase also provides link management: URLs are managed as objects, so that links point to logical pointers instead of to physical files.
Using the edition feature of the server, DynaBase tracks which editions an object has been published in, but not where an object is used. In other words, you can easily see the editions to which an object is made visible, but the administrator doesn't have a way to see all of those in context.
A typical document-management system also has either an integrated workflow tool or hooks to incorporate third-party ones. DynaBase, for now, has hooks for other workflow tools.
The DynaBase server is now client-aware, meaning it recognizes the name of the browser captured by the Web server and can process pages accordingly, if the publisher decides to offer different editions for different browsers.
Windows client. The staff contributing to the site interacts with DynaBase through a Windows client, which works locally or remotely over tcp/ip connections. It has a nice Explorer-like interface for checking files in and out, adding new content and modifying attributes.
Administrators, production managers and programmers will appreciate the user interface Inso has put on the configuration-management functions. When you add a new document class, you specify its tags (by loading and editing an existing set or setting them out from scratch).
For most people today, that means loading a flavor of HTML. But because DynaBase treats HTML as an SGML document type definition (dtd), the system has the foundation for xml support built in from the ground up.
For each tag, you define its attributes, which are then available for searching. Any of the tags can have scripts attached, and the attachment can be made global or specific to an edition.
This is one area where Inso has done a nice job with the user interface. Programmers still write scripts in a scripting language, but the DynaBase client interface makes it much easier to keep track of scripts and how they are used, without embedding them directly into the documents.
Some might be put off by the lack of a Mac client; we may see a Java client before a Mac-specific one.
How does DynaBase differ from other content-management systems? There are several features worth noting:
- Tag-aware indexing. Most full-text indexing engines can ignore tags, but few are making use of them. DynaBase, picking up the full-text engine of DynaText, recognizes tags. That feature will become increasingly important as publishers begin adding unique tags to their files. These unique tags are not specific to DynaBase-they can be whatever tags you like, and the system is open to whatever behavior you attach to them. Obviously, Inso is in an enviable position if xml takes off, because it already has a system designed from the outset for handling a variety of document definitions.
- Parsing and trapping on tags. It's common with middleware products today to add server-specific tags in your html documents in order to invoke specific actions, such as querying a database. DynaBase is more open and more flexible. The data server not only allows you to attach scripts to any element in the document, but also provides an interface for defining the context for invoking the scripts.
For example, suppose you author a set of documents with a subset of the HTML 3.2 spec. You define that, for a specific edition of the site, certain scripts run at the start or end of H1 and H2 tags. But you can set up different editions that map different scripts to the same tags-for example, inserting a reference to a css style sheet or embedding formatting directly in the HTML. Visitors can be presented with two completely different views based on the one document, which was authored using straight HTML instead of by embedding server-specific tags. Working with document classes. DynaBase can handle different tagsets, treating each as a document class with its own document type definition. This makes it easy to associate styles and scripts with different types of tags and document classes. The server also can validate documents against classes.
There are several things to like about this approach. One, it leaves the system open to a variety of authoring tools. Two, it makes it much easier to separate the tasks of authoring, design and programmatic layout (where layout means the navigation views and behavior, typically achieved through programming scripts).
- Validation. Once you've set up a document class, DynaBase will check incoming documents to make sure they conform to the tag and structure you defined. Ones that don't can be kicked back to the author for fixing, before they get posted to the public.
- VB scripting. The scripting language is Visual Basic, in contrast with TCL, which Vignette uses. For a general-purpose audience, we expect that Visual Basic will have more appeal.
- Page assembly can be dynamic by attribute, or by a place in the hierarchy. Quite a few systems will build dynamic pages based on attributes; few will build documents based on the results of parsing a tree (get next child, get parent, etc.).
Customers. Among the first customers for DynaBase are Digital Consulting, Inc., the Internet Services Group of Thomson Publishing and US Web. DCI uses it to run its Web site of trade show and seminar information (Internet Expo). The site has more than 1,000 documents and allows external vendors to operate as part of the content-development team. Thomson's installation is more ambitious, incorporating hundreds of different Web sites and 250 separate authors in a shared hosting spanning 34 separate publishers. Thomson's services group wanted to retain centralized control of technology while allowing content development to be distributed to the individual publishers; Inso claims that with DynaBase as the infrastructure, Thomson plans to grow the system by a factor of ten this year. US Web, a well-known Internet service provider, is using DynaBase to provide secure, remote access for 65 nationwide affiliates. Among the projects that have been developed is the duPont Registry, a Web version of a classified advertising magazine. The site is managed by the client in Florida but hosted at US Web's corporate site in California.