Women in Technology

Hear us Roar



Article:
  The State of JAXB: Availability, Suitability, Analysis, and Architecture
Subject:   Senior Architects?
Date:   2004-05-10 13:25:13
From:   chudak
I'm not sure why an article titled "The State of JAXB" was essentially a Java vs .NET shootout regarding XML binding.


I'm also wondering if the author done much development of enterprise applications or if he understands his target audience (whom he claims are 'senior architects').


He writes "this approach has the disadvantage of requiring knowledge of Ant...". Every project that I have ever worked on used ant for builds and deployments. I would hope that anyone with the title of senior architect or senior developer for java based applications has at least a working, if not an expert understanding of Ant.


He also complains about JAXB creating interfaces instead of concrete classes. Again, anyone familiar with sofware engineering, especially in OO languages like java will tell you "program to the interface". There is no reason for the client code to know about the concrete implementations. You can easily externalize the object creation to a delgate factory to allow changing JAXB implementors very easy instead of hardcoding calls to ObjectFactory everwhere.


Also, his 'discomfort' with JAXB not having constructors for required attributes is curious. As JAXB objects are implicitly java beans, anything but a no-arg constructor is pointless if you want to interact with your beans reflectively and would 'muddy' the waters, not to mention break the idea of programming to the interface.


FWIW, I've used JAXB on a couple of projects and there are some serious limitations with Sun's implementation but, in general, it does solve a common problem that we've encountered and is a large step in the right direction.

Main Topics Oldest First

Showing messages 1 through 1 of 1.

  • Senior Architects?
    2004-05-11 10:39:32  Michaelok [View]

    Agreed. The unfamiliarity with Ant and basic Gang of Four Patterns, so common place in Java nowadays, leads me to wonder about the rest of the article. More of a "learning experience" article, switching from Microsoft to Java?

    Both Castor (and now JAXB) are proven tools - if you read the Castor mailing list - there isn't much you couldn't do with Castor, and if you did find a scenario, the developers were quick to have a workaround or fix.

    That's the reason for the extra complexity of these tools (JAXB was in beta for over 3 years), but what you get is incredible power such as JAXB's support for multiple implementations. Downright impressive.

    Sure, .NET's version might be simpler out of the box, but no doubt you'll run into a wall quickly, just like the rest of that toolset.