Review: OCR App Follows The Rules

What Taskmaster brings to Datacap solution providers is a tremendous amount of flexibility, which translates into value-added services revenue through customization.

Its OCR process is decoupled from its processing and recognition rules. In addition, Taskmaster uses a distributed architecture to scan and process documents, so rules can be either hosted on servers as Dynamic Link Libraries or run as Web services. Users also can combine the two approaches as a single client/server solution.

Taskmaster's distributed architecture allows users to perform the scanning and verification of the recognition results at remote locations.

Once a scanning process is finished, users can place the files into a network folder via a Web browser interface. The scanning software installed at remote locations communicates with the Taskmaster system.

Taskmaster's programmable environment allows partners to control the entire OCR process life cycle. Processes for data validation, release of data, page identification and exporting of data all run in rules and can be controlled that way. Taskmaster is completely programmable, so it works in other environments.

All of Taskmaster's rules are designed, developed and tested in Datacap Studio, which combines many features of a modern IDE. Solution providers will find Studio's layout and its building blocks for identifying and matching characters quite intuitive.

CRN Test Center engineers found the recognition process straightforward, testing the software using Hewlett-Packard's Scanjet N6010 (see review).

After scanning a form into an image and placing it in a folder, engineers were able to quickly build a set of rules to recognize the fields. Taskmaster's Fingerprint matching technology adapts to virtually any scanner. The software is tolerant to most image variations.

Taskmaster does not impose limits in the way rules can be applied—be it to entire batch runs, individual documents, a page or a field on a page. Once rules are coded, solution providers can deploy the same files in the Rulerunner service, enabling partners to test a page, perform OCR on it and get exactly the same results in production.

It only takes four steps to setup a Rulerunner Web service, which is written in ASP.Net 2.0. By simply pointing to the service folder using a virtual directory in Internet Information Services (IIS) and starting a Taskmaster Server service on the same Web server, partners can immediately test live data.

Other OCR vendors use more traditional client/server configurations, which lock users into a LAN environment. Since testing documents in a live environment can be difficult, others prefer to emulate the environment by using a setup program and sample documents to configure the recognition and data capture process.

However, setup environments usually are not executed the same way as production environments, so customers can get different results in live usage.

Taskmaster's power comes to light when processing complex invoices and forms. The software eliminates the need to rely on separator pages when scanning forms in batches. It can recognize pages automatically using rules that identify fields in each page.

Datacap's partner program requires minimum sales of $67,000 to receive training. Margins range from 25 percent to 40 percent.