Why Java won't be replaced in a hurry.
May. 30th, 2008 01:21 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
mundensNor C for that matter. Or, unfortunately, Visual Basic. Recently, certain posts by others have made me want to write a diatribe pointing out why old languages like Ruby and Python are unlikely to replace Java in the short term, if at all. Then I found someone had already done it for me, complete with relevant statistics and pretty graphs. (I say "old" because Java was released to the public after Python (1991) and at the same time as Ruby (1995), these languages are not "new hotness" in comparison to Java at least. )
In the comments to the post I link to above, people talk about how Ruby has Rails and Python has something similar, as if Java didn't or couldn't utilize similar tools.
As an example, a web-based application I'm currently working on has over a thousand use cases and 260-plus database tables. 95% of the code for the entire application, from the Hibernate layer data access through to the Spring-based UI with AJAX support, is generated directly from the UML object model. That's somewhere in the region of five thousand application-specific source code files, if you include all the JSP's and .js scripts.
Now admittedly we didn't develop this tool until I'd seen Ruby on Rails and demonstrated it to a few people and asked the question "Why the fuck aren't we doing this?", so Rails and similar things are responsible for getting us thinking about what we were doing. But it took one of our guys a couple of months to assemble the tool-set primarily utilizing existing open source technology with a bit of customization for our own purposes. I find it hard to believe that we were the first to do this for Java! I mean, the idea of innovation coming from this place? C'mon, that would never happen, right? :)
We're now looking at delivering an application using three people over about six months, which would previously have been estimated at something in the region of twenty person-years. Which is probably bad news if you're looking for a job in the industry, as we now need less people, but is great news for our customers and the company, because it's costing them a hell of a lot less that it would have, and makes us an order of magnitude more competitive than we were. All we gotta do now is try and figure out a way of reducing the number of source code files...
In the comments to the post I link to above, people talk about how Ruby has Rails and Python has something similar, as if Java didn't or couldn't utilize similar tools.
As an example, a web-based application I'm currently working on has over a thousand use cases and 260-plus database tables. 95% of the code for the entire application, from the Hibernate layer data access through to the Spring-based UI with AJAX support, is generated directly from the UML object model. That's somewhere in the region of five thousand application-specific source code files, if you include all the JSP's and .js scripts.
Now admittedly we didn't develop this tool until I'd seen Ruby on Rails and demonstrated it to a few people and asked the question "Why the fuck aren't we doing this?", so Rails and similar things are responsible for getting us thinking about what we were doing. But it took one of our guys a couple of months to assemble the tool-set primarily utilizing existing open source technology with a bit of customization for our own purposes. I find it hard to believe that we were the first to do this for Java! I mean, the idea of innovation coming from this place? C'mon, that would never happen, right? :)
We're now looking at delivering an application using three people over about six months, which would previously have been estimated at something in the region of twenty person-years. Which is probably bad news if you're looking for a job in the industry, as we now need less people, but is great news for our customers and the company, because it's costing them a hell of a lot less that it would have, and makes us an order of magnitude more competitive than we were. All we gotta do now is try and figure out a way of reducing the number of source code files...
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
Code generation from UML
Date: 2008-05-30 11:53 pm (UTC)MDA was first picked up by the OMG back in 2001, if memory serves, and now covers everything from analysis, to code generation, to automated testing, and reverse engineering.
There are loads of tools that support it - including Eclipse.
I completely agree with you about Ruby and Python "replacing" Java - I read the same stories and just about lost my coffee all over the screen. Heck, I still know some large systems that are still in Cobol!
"Real programmers never die, they are simply cast into (void*)"