In-Depth

BIRT: Users Optimistic about Emerging Reporting Engine

Based on the experiences of several users, BIRT looks like a good start—which should only get better

• By Stephen Swoyer
• 06/15/2005

Last week, the Eclipse Foundation delivered a version 1.0 release of its Business Intelligence Reporting Tool (BIRT), a Java reporting tool for the Eclipse IDE.

At this point, BIRT is admittedly a mere blip on the BI reporting tools radar screen. Nevertheless, few dispute the pressing need for a standard Java-based reporting engine for J2EE applications. Based on the experiences of several users, BIRT looks to be the best of the bunch—and should only get better.

Chris Downey is a report developer working in the government sector. Like many users, he stumbled upon BIRT while searching for a usable J2EE reporting engine. Like many BIRT adopters, he admits that BIRT 1.0 isn’t without its shortcomings.

At the same time, Downey stresses, it shows a lot of promise. While he’s not ready (or able) to deploy BIRT in place of his existing reporting solution, InetSoft’s StyleReports, Downey plans to keep his eye on the Eclipse BIRT project. “[StyleReports’ lack of] RTF support in text fields is a huge problem. Unresponsive, relatively unskilled technical support is another issue that has many of us annoyed,” he says, citing reflexive advice to “upgrade to the latest version” of StyleReports. “[W]e … are doing everything possible not to pay a single dollar more to InetSoft because of our negative support experiences,”

Of course, Downey concedes, BIRT 1.0 is by no means a slam dunk. “I’ve experimented just enough to know that BIRT is not ready for our needs yet. It’s still quite raw. It integrates nicely into Eclipse, which is a big plus,” he explains. “[W]e will probably use StyleReports for another year or two because it will take time for BIRT to evolve and for me to enroll management into switching over to BIRT instead of just ‘plugging holes’ in BIRT.”

Nevertheless, Downey stresses, he’s encouraged by the BIRT project, which—as he sees it—isn’t just about developing a Java-based reporting plug-in for Eclipse, but (instead) about defining an open standard for J2EE reporting. “I’m very optimistic, but my timeframe is years, not months. BIRT does have the potential to become a ‘one-stop shop’ as well as an open standard for J2EE reporting. I intend to become actively involved in the project.”

Martin Miguel López, a programmer with Argentine services and software vendor TopGroup, also sings the praises of BIRT, which—like Downey—he discovered while searching for a better alternative to his existing Java-based reporting tool.

“Our first tests of BIRT have returned good results, basically because it seems to be a good solution to our current report tool problems,” he says.

TopGroup is currently using JasperSoft’s Jasper, says López, but has been frustrated by that product’s shortcomings, which include a steep learning curve for junior report developers, as well as performance issues. “We have a strong open-source orientation, because the economic issues of our country make our enterprise less competitive if we choose commercial solutions. So this means our (serious) report tool choices are minimal.”

Like Downey and other users, López has an optimistic take on BIRT, which, he concedes, isn’t quite a silver bullet—at least in its present incarnation. “We feel very optimistic about BIRT, as it has become very mature in a very short period since it was proposed as an Eclipse project,” he says.

Even programmers who aren’t necessarily working with BIRT at least have it on their radar screens. Take Anthony Raj, a Crystal developer with an Indian software vendor. He says his firm will almost certainly adopt a Java reporting tool—it’s just not sure which one. “[It] helps in marketing, cross platform, and some of these things are enterprise ready. In fact, we've been having an itch to compare these tools with the features that [Crystal Reports] dominates and helps enable those features over a period of time,” he explains.

BIRT is currently third on Raj’s list, behind Jasper Reports and DataVision, an open-source reporting tool that’s described as similar to Crystal Reports. As it matures, however, it could become a more viable solution for Raj and other developers. That’s the hope of most users, anyway.

For enthusiasts like Downey, the BIRT project is important for another reason, too. “Another problem is that many J2EE architects and developers tend to treat reports as throwaways and put very little thought into designing a report as a specialized program that will need to be maintained,” he observes. “Our reports have tons of difficult-to-find scripting that attempts to mimic the application data, and over the years this has grown into a second code base that is a nightmare to maintain. I’m looking at BIRT as a way to push changing this mindset and making reports more technically interesting to work on.”