FRAMINGHAM (04/03/2000) - Setting up an internal Web site is a logical way to give corporate users a single point of entry into enterprisewide repositories of information. Unfortunately, merging multiple repositories of documents in different formats into one searchable index is cumbersome and time-consuming.
Enter LivePublish Suite 2.0 - NextPage LC's XML-based tool for creating collections of documents that can be searched and navigated from any standard Web browser. This product provides full-text indexing for a host of file formats - anything from Microsoft Word to Adobe Acrobat PDF files - and lets Web administrators assign access privileges and tailor content based on a user profile. It also offers e-commerce capabilities, supporting pay-per-view and subscription-based business models.
Overall, we found LivePublish Suite 2.0 to be quite powerful. It has the potential to dramatically reduce the time it takes to publish large repositories of documents on the Web. However, the product has some annoying quirks, namely an antiquated command-line interface for updating existing document collections.
The product is designed around a site model, which means that each LivePublish site contains one or more hierarchical collections of documents. Document collections are composed of multiple documents that get displayed one at a time.
LivePublish Suite 2.0 ships with several components, including LivePublish Builder, Server, ToolKit and Distribution Kit. The Builder component is used to organize the collections. Builder includes production tools that index, validate and incrementally update content that is then served up by a LivePublish Server. For backward compatibility, NextPage LC provides a migration tool for users who want to pull data from existing Folio 3.x and 4.x infobases into the LivePublish collections. Folio is NextPage's precursor to LivePublish.
LivePublish Server works with your existing Web server to display content to end users. Under Windows NT, LivePublish Server will peacefully coexist with Microsoft Internet Information Server 4.0 or Netscape Enterprise Server 3.62.
LivePublish Server uses a stateless, multithreaded engine to serve up documents. This enables scalability without requiring significant hardware resources. Like other NT-based Web server products, LivePublish will perform best on a system with specifications that far exceed the minimum requirements of 96M bytes of RAM and 50M bytes of hard disk space. For large-scale implementations, it would make sense to dedicate one or more systems to serve up LivePublish content.
The LivePublish ToolKit is a set of APIs used to extend the product to perform functions such as creating a custom component that modifies the way information is displayed by the LivePublish Server. The LivePublish Distribution Kit is a tool for publishing content to CD-ROMs.
Document collections are created and updated via LivePublish Builder's lpBuild utility. Surprisingly, lpBuild is command-line driven, using XML-based content definition "makefiles" to assemble the document collections. Makefiles specify the structure and content of a collection. In today's graphical user interface-based environments, using a tool that harkens back to the days of MS-DOS can be a turnoff to some organizations.
Nevertheless, when you get acquainted with lpBuild's command-line syntax, it is possible to build a document collection without any additional programming or script writing. After using the product for a while, we found that the key to creating error-free document collections is a clear understanding of the proper structure of a content definition makefile. If you're not familiar with XML or Standard Generalized Markup Language, expect a learning curve here and be prepared to rewrite definition files as errors crop up in the build process.
LivePublish users would benefit from a robust set of debugging tools, which are lacking. To debug syntax errors, you'll need to turn to an external XML validation tool, such as XML Pro or XMetaL. LivePublish Builder provides routine logging capabilities, but they're not adequate for publishing complex sets of documents.
Once we understood the Document Type Definition used in the makefile, we found it easy to map document characteristics, such as header and paragraph formatting information, on a set of Word and Excel documents to their Web-page formatting counterparts.
Document collections are added to the LivePublish server through LivePublish Site Administrator, a Windows-based application. Through the Site Administrator, you can password-protect documents and mount them from network drives. Site Administrator runs locally or remotely, offering the same functionality in either mode. To prevent unauthorized access, a username and password are required before performing any remote administrative tasks.
Each document collection is mapped to a series of display templates.
LivePublish Server ships with templates for publishing information in English, Spanish, French, German, Dutch and Portuguese. All templates can be modified but will require you to have some advanced knowledge of HTML.
Installing LivePublish Suite 2.0 on the server was straightforward. The software's install detects the resident Web server, installs the appropriate files and establishes directory aliases and configuration file modifications.
From an end-user perspective, a LivePublish site is easy to navigate. Users can view information in an HTML frames or no-frames environment. Different views provide access to search tools, search results, open documents and a directory tree of available information. The utility of the user interface really hits home when you use the search tools. LivePublish Suite 2.0 highlights search hits in target documents so you don't have to visually search a document for key words. An available search hit list displays a brief summary of each search hit and links to the corresponding document. Advanced search options include Boolean, Keyword in Context, phrase and proximity searching.
Despite some shortcomings, LivePublish Suite 2.0 represents a solid option for organizations looking to publish large document collections to the Web. It is fast, and its use of emerging XML standards should leave you in good standing.
Coopee is the technical director at Industrial Media, a consultancy in Ottawa.
He can be reached at email@example.com.