InformationWeek.com

Welcome Guest. | Log In| Register | Membership Benefits

March 13, 2000

For Some, a Solid Knowledge-Management Foundation
ZyLab's suite can be a good basis for companies that want to use existing knowledge better

By Jeff Angus

Related links from our sister publications:

Computer Reseller News Stopping knowledge overflow (2/28/00)

InternetWeek Grand Vision, Tall Order (1/24/00)

TechWeb Knowledge Management Gains Sharper Focus (1/17/00)

TechEncyclopedia

Need a definition of a technology term? Look it up here:

Send Us Your Feedback

t's pretty rare that a product, even a suite of products, comes along that we think makes for a good foundation for knowledge management. There are so many processes, people and products required to deliver an adequate solution that no one vendor is going to provide much more than a piece, let alone a whole foundation.

I recently tested a suite that could be a stellar foundation for companies that want to make better use of existing knowledge rather than to redesign themselves to be a fully functioning knowledge organization. The product is Zylab International Inc.'s ZyImage 3.0.

Zylab is one of the longest-running acts in search technology, and certainly the search component of ZyImage is top-notch. But Zylab has managed to blend in utilitarian document-management and imaging capabilities. The result is a suite that balances features with ease of use that can be a productivity enhancer for companies whose goal is to boost their recall.

ZyImage 3.0 is based on a series of modules you use in a structured workflow. You can automate the workflow, execute one or more modules manually, or both. If you have workstation licenses, you can distribute work, for example, mass batch-scanning of standard forms and paperwork through a central facility while making it possible for individual users to contribute unstructured documents from their desktops. Its completeness and flexibility mandates the need for training and solid preplanning by the IT group.

The addition of ZyAlert in this version adds some push capabilities to the product, strengthening its status as a knowledge-management foundation.

The panoply of modules in ZyImage is collected into three consoles: one each for input, process, and output.

ZyScan is the workflow-sequenced program for input of all kinds of files into discrete directories. There's the ZyScan module for acquiring scanner input and ZyImport for targeting files on disk for acquisition. ZyField is the module for assigning what properties you want to track (such as author, date, subject, and importance). ZyOCR converts images of documents into text.

The workflow is logically arranged and linear. Users won't always need to exercise all the steps (you might only have scanned pages or only existing disk files). The program's interface is flexible, allowing you to skip unneeded steps while still alerting you to sequences that might be incomplete.

The system supports "jobs," recipes you predefine for repetitive jobs and the options for these are beyond adequate for almost all the document workflows one would need to launch.

The processing program is ZyIndex. The main module is first in the flow: ZyBuild, in which you specify files to be included in an index. This is again flexible, letting you exclude specified file types from the index. You can use ZyTimer to schedule updates or maintenance drills automatically. There are tools for publishing to laser disk or CD-R media. ZyWeb publishes an interface between the index and HTML, a form that generates search pages for the indexed content so Webmasters can quickly deploy a browser client access to stored documents. The generated pages are formed from included templates specific to the ZyWeb server, but you can modify the templates (or create your own) to give them the ability to work on other Web servers. The program includes an encryption module for users who want to encode content.

I found the building process mostly very speedy, even on 100 Mbytes of files. There are some tricks to learn on what kinds of files might need different levels of ZyIndex's "fuzzy search." Some HTML pages didn't index as completely as I thought they would. Others did.

The output module is ZyFind. You can use it as a Windows program from your desktop, or, if you've used the ZyWeb publish to HTML, access the index from there. ZyFind is very powerful on complex searches. The program presents a document list with "hits" in it. You can view the document on your screen either through a page viewer (that works surprisingly well within a browser) with your hits highlighted. You can also launch the document in its authoring program if you have it. I found ZyFind robust and very fast, although I blew it up on more than one occasion when it was trying to display hits in Adobe Acrobat files.

That's an interesting place for an imperfection, because in some ways, ZyImage can work as an Acrobat replacement. It delivers highly graphical images of pages but with full access to text, Acrobat's claim to fame. While Adobe's product was originally designed for print output, the format is popular for high-resolution Web content. I don't think shops that adopted Acrobat as a standard will want to replace that kind of publishing with ZyImage, but the ZyLab product can add a lot of structure to the value of PDF content by indexing it and integrating it with non-PDF files.

ZyAlert provides an interface for delineating searches and triggering E-mail notification or file attachment to specified people. This will deliver great value in specific markets where, for example, a competitive intelligence analyst will want to keep up with every reference to certain companies

Online documentation is good, not great. Some simple designs, such as sequenced help topics you can navigate by paging arrows, would make great sense given the program's linear nature.

ZyImage isn't a knowledge-management solution for most shops. But it's a well-designed foundation, and some vertical markets, such as equipment maintenance or legal services, should find this product irresistible.

Back to This Week's Issue
Send Us Your Feedback
Top of the Page