Opinion: Granular choice, low barriers to entry foster software improvement.
Does this sound to you like an accurate statement? "Developers gravitated to Java because they could focus on writing business logic. The run time handled everything else. A lot of people want to return to that level of simplicity, even for multinode, enterprise applications. Thats the goal."
Im quoting a comment made today by Ari Zilka
, CEO of Terracotta
, a San Francisco-based supplier of Java scalability technologies
. Im not saying hes wrong, and in fact I think theres a lot of useful truth in what he says, but I also think theres opportunity to be found in recognizing where thats not the whole story.
For one thing, the Java run time may "handle everything else" in subtly but importantly different ways on different platforms. Key application behaviors such as fulfillment of real-time responsiveness requirements
may depend on behaviors that can vary considerably while still remaining within the realm of Java specification conformance. Its essential to know the difference between whats allowed, whats required, and what can be made to happen by saying explicitly in the code that you want it.
It might seem that the latter course clearly takes you away from just writing business logic, but perhaps it begs the question to say that: When does code that controls an applications run-time behavior cross the line from business logic to process implementation?
The question of what is the scope of an enterprise developers concerns, and whats in the realm of middleware or systems software design and implementation, takes on new importance as extensible platforms emerge to let developer organizations configure their own tool sets. NetBeans
can be extended with plug-ins
that span the application life cycle from initial modeling through database architecture, GUI development and performance profiling. Standards-based interaction models like AJAX are growing into deeper platform models
, with conveniently accessible mechanisms for key functions like persistence of personalization settings, as seen in frameworks like Microsofts Atlas
, whose developer convention starts today
in Santa Clara, has active projects
under way addressing such interests as business intelligence, service-oriented architectures and application lifecycle management
Im seeing the kind of third-party toolmaker interest around this weeks EclipseCon that used to distinguish events like the Borland or Symantec developer conferences. (At least the Software Development shows
still go on, with highlights like the product awards
named in honor of the hypercaffeinated soft drinks that legendarily propel
developers late-night sessions.) We can expect to see some plain and fancy positioning of various toolmakers, including Borland
, as they vie for the honor of being first among equals in adding value to the Eclipse effort.
Whats key is that so many flavors of specific expertise can now find their way to developers in the approachable, integratable form of low-cost but full-functioned environments. Those experts can now enter the toolmaking market more easily, which is good--and the users of those tools face more granular choices about the tools theyll adopt, and correspondingly the skills theyll develop and the strengths theyll market to customers, which is challenging but certainly not bad.
Tell me what youll be hoping to hear from EclipseCon this week in particular, and from development toolmakers this year in general, at email@example.com
Check out eWEEK.coms for the latest news, reviews and analysis in programming environments and developer tools.