JAVA IDE TOOLS SPECIAL REPORT
- By John K. Waters
- September 1, 2005
The Java tools market is getting an extreme makeover, and like the nipped-and-tucked
subjects of a TV-reality show, the commercial vendors are feeling some pain,
bracing for an uncertain outcome and hoping for the best when the bandages finally
come off. What’s more, these changes are more than skin deep. The catalyst
of this transformation is Eclipse and its commoditizing effect on the basic
components of the integrated development environment. Thanks to the influence
of this open-source, platform-agnostic tooling platform, the new face of the
Java IDE will bear a closer resemblance to a lean architected plugin environment
than the traditional feature-stuffed tool suite (which, let’s face it,
could use a little liposuction).
“The traditional IDE has had its day,” says Forrester Research
analyst Carl Zetie. “For the last decade, the IDE vendors have been so
busy competing with each other on features and functions they lost sight of
what customers actually wanted, which was something much simpler, much lighter
and much easier to use.”
Zetie cites a recent Forrester survey of Eclipse users, which found that most
are combining an average of three to five plugins with the basic Eclipse software
development kit to create their preferred development environments. “The
list of features in something like a Borland JBuilder, an IBM Rational Application
Developer or an Oracle JDeveloper is equivalent to Eclipse plus maybe 30 to
40 plug-ins,” he says.
To be fair, Zetie adds, the commercial IDE vendors are in a tough position.
The need to appeal to as many customers as they can requires they cram as many
features as possible into their development environments. Although developers
may value only 10 to 20 percent of those features, which 10 to 20 percent varies
from developer to developer.
“Every time one vendor came out with a new feature, all the others had
to react and do the same thing,” observes Gartner analyst Mark Driver.
“The result was feature creep and bloating. There’s no doubt that
Eclipse is changing that model, basically doing what opens-source products do:
commoditize the lowest common-denominator functionality and create a fundamental
consolidation. The smart IDE vendors will stand on that foundation and focus
on their true value-add without getting mired in the issues of building yet
another debugger and text editor when there’s no longer any real value
To varying degrees, all the major Java IDE vendors are gearing up for this
new platform-based tooling model. Naturally, IBM’s Rational Application
Developer for WebSphere is built on Eclipse. The Eclipse open-source project
grew out of technology first developed by IBM and later open-sourced.
Borland Software, one of the founding members of the Eclipse consortium, recently
amped up its support for the tools platform, even as its chief evangelist, David
Intersimone, declared it “the commercial IDE killer.” The toolmaker,
best known for its JBuilder Java IDE, ships several products based on Eclipse,
including Together Control Center.
BEA Systems’ Project Pollinate is an Eclipse-based development environment
and toolset designed to integrate with Apache Beehive, the application framework
from its WebLogic Workshop Java IDE it donated to the Apache Software Foundation.
BEA is now leading the Eclipse Foundation’s Web Tools Platform project.
SAP’s NetWeaver Developer Studio now sits on top of the Eclipse infrastructure
and integrates a range of functions through plug-ins. Macromedia is working
on an Eclipse-based, next-generation, rich-Internet application development tool,
codenamed Zorn. And Compuware has committed to building all future versions
of its OptimalJ on the Eclipse framework.
Sun Microsystems is the obvious exception, because of its technological (and
ideological) commitment to the open-source NetBeans IDE and platform, which
has its own devoted following. The netbeans.org project claims more than 5 million
downloads of the free Java development environment. “NetBeans out of the
box is actually a more full-fledged and fully rounded IDE that the Eclipse JDK,”
says Driver. In addition, NetBeans is free for both commercial and non-commercial
Oracle, though on board in spirit, will be a limited Eclipse supporter for
the near future, because its JDeveloper IDE is so tightly tied to its Fusion
Middleware offering. The company is currently spearheading a JavaServer Faces
tooling project within the Eclipse Foundation and expects to work on an Eclipse-based
infrastructure over the next 3 or 4 years. In the meantime, Oracle has made
the headline-grabbing decision to give its JDeveloper 10g Java IDE to any developer
who wants it.
Oracle’s JDeveloper strategy underscores the radical nature of this shift
in the commercial IDE market. “To succeed in this new market, the vendors
will have to focus on those pieces that have not become commoditized,”
Driver says, “those higher end services—what SAP is doing, for example,
with process-driven development—or what IBM is doing with integration
across all of their product lines. Vendors that don’t do that are competing
against free. Whenever you compete against free, sooner or later, you’re
going to lose.”
Although not its original purpose, Eclipse spawned its own free IDE, which
is fast becoming the most widely used Java development environment. Many analysts
consider it to be on track to become the de facto alternative to Microsoft’s
Visual Studio.NET. Last year, industry researchers at Evans Data declared it
the most popular Java IDE in North America, with strong growth in Europe, the
Middle East, Africa and Asia-Pacific.
The Eclipse SDK combines components from the Eclipse Project, bundled into
a single download. They include the core frameworks and services on which plug-in
extensions are created (the Eclipse Platform), the plug-ins that implement a
Java IDE (Java Development Tools, or JDT) and a group of views and editors for
building Eclipse plugins (the Plug-in Development Environment, or PDE).
“The guys who built the JDT were actually developing something much harder,”
says Mike Milinkovich, executive director of the Eclipse Foundation. “They
were creating a framework for building Java development tools, and they built
the JDT as an example, to show that it could be done. What we’re trying
to do with all of our projects is to create a set of frameworks that developers
can use and extend for their own application domain, or vendors can extend to
build their own products.”
All of this change is good news for Java developers, who are now entering what
promises to be a unique buyer’s market. Among other things, the commoditization
wrought by Eclipse should lower barriers to market entry. Smaller toolmakers
will no longer have to come up with the resources to reinvent another Java debugger,
test editor or project management workbench. The result: more choice for developers.
“In the past, those companies have had to pitch to a very small market
and people who were willing to buy into their entire product,” Zetie says.
“If you were trying to sell, for example, a full-featured UML modeling
tool, somebody had to be willing to ante up thousands of dollars to buy into
your entire vision of what UML modeling should be. The new commoditized model
is all about adding value. You could produce, say, the world’s best use-case
editor, and somebody could take that up and use it with their preferred interaction
diagram editor and their preferred physical class modeler. And you could sell
it for hundreds instead of thousands.”
But the value-add is a moving target, says Zetie. This year’s unique
feature could be commoditized next year. “Eclipse inevitably is going
to do for most other lifecycle phases some of what it’s done to the IDE,”
he says. “The thing about Eclipse is that it is both commoditization and
innovation in the same package. Most open-source projects are one or the other.
Apache and Linux: commoditization; Spring: innovation. Eclipse is both at the
So, it’s going to commoditize the rest of the lifecycle the same way
that it’s already commoditized the plain-vanilla editing of Java code.”
That’s going to be tough on vendors, but it’s more good news for
buyers. On such shifting sands, vendors just can’t establish the lock-in
they once had. That’s the good news for the tool users: more choice and
For developers and IT managers, there’s a lot to consider in this brave
new commoditized world. Driver’s advice: Evaluate your IDE within the
context of your organization’s infrastructure, middleware and other vendor
commitments, as well as the direction of your current tools vendor.
“If you’re a heavy SAP shop, for example, you’re certainly
going to be looking at Eclipse via NetWeaver,” he says. “If you’re
a heavy Oracle shop, then JDeveloper may still make perfect sense for you. If
you’re an IBM shop, there are tons of things in RAD 6 that aren’t
in Eclipse—things like modeling support and business process management
tools. It’s about using the right tool for the right job.”
Another consideration: the hidden costs of freebies. “The problem that
we find now is that people expect the Eclipse JDK to be free,” says Driver.
“Well, it’s not free—not after you take into account the time
it takes to do the best-of breed, plug all the stuff together and get everything
Also, Eclipse SDK is a very code-centric IDE, which can add to its cost. It’s
not the kind of product you can throw at 4GL-style developers—the developers
targeted by Sun’s Studio Creator or NetWeaver. And Eclipse still has a
strong self-service flavor; you can’t call for service and support for
the free stuff.”
Milinkovich rightly insists that Eclipse is not just about Java (there are
projects currently under way to create Eclipse-based IDEs for C/C++ and COBOL
development, and another to support aspect-oriented software development with
AspectJ), or IDEs (the 3.1 release of the Eclipse platform covers Web tools,
business intelligence, reporting, testing and performance). And he says it was
never about competing with toolmakers.
“There’s always lots of room to build differentiating product functionality
on top of the Eclipse framework,” he says.
Sidebar: The Tools Market
is Changing, but Demand for Support Isn't