In the last Pro-Spective we detailed the benefits of IBM's Workplace Web Content Management (WebCM) solution and its powerful integration with IBM's WebSphere Portal Server. Many of you may already have another Web content management solution but are interested in migrating to WebCM to benefit from the additional value. This article details the repeatable steps involved in migrating from Vignette to WebCM, and a redeployment strategy.

Since Vignette is a database-driven product, the advantage of its open DB schema is that it gives you an easy way to extract data. After gathering the required information on Vignette database tables and fields, data (such as navigational categories and content of different types) can be analyzed and easily mapped into the site structure and content objects in the WebCM Repository.

By analyzing the Vignette CMA (Content Management Application), you can:

• Identify Vignette content types with their fields and their mapping into corresponding WebCM content templates that will be re-created using the WebCM authoring environment (an easy and quick process).
• Identify content categories - other than navigational - that will be re-created using the WebCM authoring environment as taxonomy(s) with categories.

Analyzing the Vignette workflows provides the information needed to recreate corresponding workflows in WebCM. The corresponding Vignette database tables provide information about each content workflow status.

Repeatable Migration Process
The data migration process is a repeatable process and can be automated by creating a migration tool. One way to implement the tool is as a JSP with an actions menu for the various processes, such as export navigational categories, export content, etc. Keep in mind that the migration tool needs to be deployed on the same WebSphere Portal server instance where WebCM is installed.

With a migration tool, it's possible to create very basic actions and configure them to do particular export tasks. For each action the tool can provide a corresponding JSP that will be displayed to the user to gather some input, if any is required, to initiate an export task, and to display any output (logging/warnings/error message).

The tool basically needs to do three tasks; their implementation can be located in separate modules (JAR files).

EXPORT FROM VIGNETTE
Using SQL queries, access data from corresponding tables in the Vignette database. For this task, a configuration file can be used to specify Vignette content tables and fields per content type, so that SQL can be built dynamically, not hard-coded; or it can specify complete SQL queries.

The tool can provide some basic export tasks (content export, categories export); the list can be extended as needed.

TRANSFORM
Transform extracted data into appropriate intermediate (Transformation) objects. Transformation objects have fields corresponding to a given content type (content template) fields, workflow status, and content profile (categories and keywords), and provide set/get methods to them. These objects may have to be created manually for each content type and provided in a separate JAR file that's shared by the export-import process.

Transformation objects are used by the Export task. While accessing the Vignette database, data get put into transformation objects and kept in a Vector object that's passed into one of the Import task methods.

A Hashtable object can be used as a Transformation object as well. With this approach Hashtable object keys will correspond to either:

• The WebCM content template fields for content migration
• The site structure parent-child relationships for site structure migration

Using Hashtable objects will allow the migration tool to be configured. Only one set of configurable variables will need to be used for content template fields and Hashtable keys.

IMPORT INTO WEBCM
Finally, using the WebCM published API, create corresponding changes in the WebCM repository. This module will provide different methods to do appropriate actions (site structure, content migration, etc.).

• Based on a corresponding workflow status retrieved from the Vignette database and stored in the transformation object, a content item will be placed into an appropriate workflow stage.
• Based on the content profile (categories and keywords) retrieved from the Vignette database and stored in the transformation object, a content item's corresponding profile properties will be set accordingly.

This module can use a configuration file to specify WebCM object names: workflow with stages, authoring template(s), presentation template(s), content field names that will need to be accessed during the import process.

These objects have to be created manually in WebCM before running the migration process, since the published API doesn't let you create them, only access them.

Figures 1 and 2 show the process flows for site framework and content migration.

The following analysis steps provide additional information for the migration process:

• Analyzing the Vignette Content Delivery Application (CDA) lets you identify content delivery components and their functionality, which will be re-created in WebCM as various generic library components and presentation templates. This is a manual process that can still re-use existing presentation-level HTML to speed up the process, unless the Web site needs a new look-and-feel.
• Sometimes as a result of tailoring Vignette applications to the needs of a particular client, custom tags can be introduced; for example, Vignette tags that build links to other content items. Since each custom Vignette tag renders certain markup, which is included in a Web page, the solution for each custom tag can be easily built based on using corresponding WebCM tags, i.e., AptrixCmpnt with start and end attributes.

To build an appropriate WebCM security model, corresponding users and groups have to be created in the WebSphere Portal WMM. If there's an LDAP server in place, the WebSphere Portal WMM can be configured with LDAP.

Integration with another search engine (for example, Autonomy Server V.4) can be easily implemented using the JSP component's references on WebCM presentation templates.

Caching Options
Regardless of the migration process, you'll need a caching strategy that can be tailored to a client's Web site requirements, which may vary widely and sometimes even require additional implementation using published (or even unpublished) functionality in the WebCM API.

To implement a successful caching strategy, the first step is an analysis of the site to determine the nature of the content to be cached. Things to consider include:

• Whether the site contains static or dynamic data, or both
• Whether "Connect" tags or URL requests in the Presentation Templates or the Component Designs are used
• Whether data from external sources such as databases will be displayed