|
|
|
|
|
The last couple of articles in this space have focused on how
the widespread use and availability of Java is making it
popular for wrapping up many existing scientific software
packages into snazzy new packages. The easy portability of
Java software makes this approach nice for developers as
well, since any platform that runs a program now will still
run it after it's interfaced with Java. Naturally, as people
have been gaining access to more codes through a friendlier
user environment, they have demanded more and more. In many
circles, there is a great effort afoot to make a ubiquitous
interface that would allow scientists to interact with nearly
any piece of existing software in a given genre. The term
usually applied to such an ambitious undertaking is Problem
Solving Environment, or PSE.
The middle tier is the most amorphously defined and complicated of the three, and it's the reason that PSE's are made up of "at least" three tiers. In fact, the middleware often consists of multiple pieces itself. In general terms, though, the middle tier must be capable of "talking" to both the back end thoroughbreds and the front end beauty queens, and letting each know what the other is up to. There are many technologies present in a typical middleware product, and perhaps we'll look at them someday. But for now, just suffice it to say that the middle tier is the glue for the PSE. What we're going to focus on for now is the front end of a Problem Solving Environment. The proliferation of Java technologies has had a profound impact on the development of such products, providing common looks and feels, and nice encapsulated entities with which programmers can work. Next week, we'll look more deeply into what's going on in the world of PSE front ends.
Go To Page: 1
The copyright of the article Java and PSE's in Scientific Computing is owned by . Permission to republish Java and PSE's in print or online must be granted by the author in writing.
|
|
|
|