I went to a presentation at my local Java User’s Group last Monday on Guice. I’m a long time Spring user (and fan), so I was interested to see what benefits it had over Spring.

I came away with the impression that Guice looks pretty interesting, but its main benefit over Spring is the lack of XML. That’s okay I guess (although I personally think the case against XML is often overstated). I certainly wouldn’t be opposed to using Guice in a project based on what I saw – except for one thing….

My big issue with Guice is the fact that it appears to be invasive. Based on the presentation I saw it seemed that you need to instrument your code with @Inject annotation for the dependency resolver to provide that dependency. That’s fine in your application code, but it means wrappers seem to be required for third party libraries. However, there are some hints from the documentation that this isn’t the case (‘Observe the Java type and the optional “binding annotation” of the element to be injected‘). OTOH, ‘sometimes you can’t use annotations, and you need to manually wire up an object (a 3rd party component for example). When this comes up, you write a custom provider‘ (See http://code.google.com/p/google-guice/wiki/SpringComparison). That’s kind of confusing, and something I’ll need to investigate more.

All in all – an interesting option, and one that certainly replaces PicoContainer as the best code-centric dependency injection container.

Random stuff:

1. There's huge demand for (good) Java programmers. (Live in Adelaide, want a job & know Java? Contact me… Also, I'm still looking for an XUL person.)
2. Javascript might be the next big language, and while everyone claims they can program in Javascript what they mean is that they know how to pop-up an alert box in a browser window. That isn't programming! (Live in Adelaide and know AJAX and possible a bit of Java? Contact me….)
3. Spring is good.
4. Doing estimates for a project when there are no requirements is difficult.
5. Doing estimates for a (different) project when the functionality, resources and timeline are all fixed is very difficult.
6. Social networking.. there's just too much to write on that topic.

I've switched my feeds at my BadMagicNumber blog to be published via FeedBurner. There should be no disruption to your normal programming, although I think some aggregators will show some non-new items as new.

I switched the feeds over using a simple Servlet Filter. If anyone wants to do the same, here's the code. This works for blojsom, but you might need to modify it slightly for your own setup.

public class FeedBurnerRedirectFilter implements Filter {
	private String redirectURL;
	
	public void init(FilterConfig config) throws ServletException {
		redirectURL = config.getInitParameter("redirectURL");		
	}

	public void doFilter(ServletRequest request, ServletResponse response, FilterChain chain) 
						throws IOException, ServletException {
		String flavor = request.getParameter("flavor"); 
		
		if ("atom".equals(flavor) || "rss".equals(flavor) 
				|| "rss2".equals(flavor) || "rdf".equals(flavor) ) {			
			HttpServletRequest httpRequest = (HttpServletRequest) request;			
			HttpServletResponse httpResponse = (HttpServletResponse) response;
			
			String userAgent = httpRequest.getHeader("User-Agent");
			if (userAgent != null && userAgent.indexOf("FeedBurner") < 0) {
				/// redirect if not feedburner
				httpResponse.sendRedirect(redirectURL);
				return;				
			}
			
		}		
		chain.doFilter(request, response);		
	}

	public void destroy() {
		
	}
}