1
2
3
4

var rank= /^(?:http|https)\:\/\/(?:.+\.)*google\..+(?:\?|\&)cd\=(\d+)/.exec(document.referrer || '');
if(match) {
  alert('position = ' + rank[1]);
}

The next thing we want to do is modify our analytics tracker code to include a custom variable:

1
2
3
4
5
6
7

var rank= /^(?:http|https)\:\/\/(?:.+\.)*google\..+(?:\?|\&)cd\=(\d+)/.exec(document.referrer || '');
var _gaq = _gaq || [];
_gaq.push(['_setAccount', 'UA-XXXXXX-X']);
if(rank) {
  _gaq.push(['_setCustomVar',1,'Google_Ranking',rank[1],2]);
}
_gaq.push(['_trackPageview']);

We’ve set the custom variable to slot 1, and the last number two means that we scope it to the session.

Finally you’ll want to wrap it up in a closure to prevent polluting the global scope:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

var _gaq = _gaq || [];
(function() {
  var rank= /^(?:http|https)\:\/\/(?:.+\.)*google\..+(?:\?|\&)cd\=(\d+)/.exec(document.referrer || '');
  _gaq.push(['_setAccount', 'UA-XXXXXX-X']);
  if(rank) {
    _gaq.push(['_setCustomVar',1,'Google_Ranking',rank[1],2]);
  }
  _gaq.push(['_trackPageview']);
 
  var ga = document.createElement('script'); ga.type = 'text/javascript'; 
  ga.async = true;
  ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js';
  var s = document.getElementsByTagName('script')[0];
  s.parentNode.insertBefore(ga, s);
})();

You can view your new statistics using the Advanced segmentation feature of Google analytics, but keep in mind that it’s only possible to harvest data from the links that are clicked on.

Well nothing actually, you can have multiple identities, sign in on several websites using that identity.  And if you don't trust a third party with your login credentials it is also very easy to setup your very own OpenID server.

Than what went wrong. well, lately several companies started to claim to support OpenID. For instance, your gmail, yahoo, hyves accounts (and soon windows live) are now OpenID identities. They all provide you with an identity but none of them allows you to another identity to sign on their sites. This completely defeats the purpose. If I wanted an identity, there are already tons of sites that will provide me one. It would be far more usefull if they would actually allow me to log in using an openid.

So next I tried Hurricaine Electric which have an endpoint at the AMS-IX so that should give me better speed, or at least better ping times. They offer no custom software but instead use protocol 41. Unfortunately it seems that my linksys BEFW11S4 router doesn't let that traffic through (not even when I place my machine in the DMZ) so I looked on.

I finally ended at SixXS which also has several tunnel endpoints in the Netherlands and offer an UDP based protocol that is capable to pass through NAT.  Their software works I get very good speeds and ping times here. The only downside I could find is that you'll have to wait for a week before you can request a /48 subnet, but this should'nt be a big problem, just be patient and besides it gives you the time to figure out how you want to configure your network.
Their service works very good and I'm glad I signed up there. 

Now all I need to do is find a router that is capable of passing protocol 41 and convert my ayiya based tunnel to one that provides less overhead.

 It always irritated me that when I specified a content-type for a web page, Internet Explorer would just look at the content and say: "your wrong" and renders the page the way it thinks you intended.
If I give a html page text/plain as content-type I intend to let it show the source, not render it as html because if I wanted that I would have given it text/html as content-type.

At least in IE8 they allow us to disable the mime sniffing by appending ";authoritative=true" to the content-type header. This way it won't hurt other browsers and we are still able to provide the mime type ourselves.

All you need to do is define an interface that  (extends the Library interface and) exposes the methods in the library that you want to access:
import com.sun.jna.*;
// kernel32.dll uses the __stdcall calling convention
// Most C libraries will just extend com.sun.jna.Library
public interface Kernel32 extends StdCallLibrary {
// Method declarations, constant and structure definitions go here
}

Next you load the library using the following code:
Kernel32 kernel32Instance = (Kernel32)
    Native.loadLibrary("kernel32", Kernel32.class);

That's all there is to it, you can now access all methods you defined in the interface.
 Checkout the project page

An example on how to call the server:

function callServer() {
 var head = document.getElementsByTagName("head").item(0);
 var myTag = document.createElement("script");
 myTag.setAttribute("type", "text/javascript");
 myTag.setAttribute("src",
  "http://example.com/application/getjs/?var1=value1");
 head.appendChild(myTag);
}

The idea is that you generate the javascript that you need by dynamically add another script tag on the page.
for an excellent example of this approach you can read this javaworld article

A brief overview of the features in this release:
* Multiple ResourceInjector instance support
* TypeLoader property support
* Global resource syntax
* SWT support
* Fully modular architecture (/Core, /Swing, and /SWT)
* Documentation (both javadoc and TypeLoader specific)
* A brand new build system and instructions on how to use it

For more information about Fuse see this entry in the Romain Guy's blog. or take a look at the project page
 

For those who don't know what Archium
is, Archium is a new programming language which I worked on during my
traineeship at the university, for more information see the archium
website www.archium.net.
I personally think that one of the best features is that archium allows
you to change your application at runtime, allowing you to update your
software while it continues to run.