I wrote the pixel tracker for a big corporation, which tracks a ton of hits for many top sites.

1) Using Math.random() to generate a UUID, is not a UUID. There are plenty of JS uuid generators out there. Please use one. 2) Why are you returning a header with what is being tracked? 3) Please please please use coffeescript.

About (1), your uuid is random with a good math.random() and a sufficiently long string with high probability (using that as a technical term from theoretical cs). So I assume that your gripe is that math.random() has bad implementations.

Recall the article recently on how a deck of shuffled cards is unique in human history? Apply same logic.

3) Why use CoffeeScript in this instance? I don't see this being a valid criticism of any JavaScript project, let alone one so trivial. Comma prefixing on the other hand...

Neat, it's cool to see pixel tracking as a middleware component.

What i'm not clear on (and i'd love to hear your thoughts about) is why you'd want it as a middleware component vs a daemonized external service such as http://documentcloud.github.com/pixel-ping/ (and if you care about such things, check out the annotated source: http://documentcloud.github.com/pixel-ping/docs/pixel-ping.h... )

full disclosure: i joined DocumentCloud a few months ago.

pixel-tracker works with express out of the box and doesnt really imply that you will want to aggregate hits with a key or any certain way. It just provides information about a request you can use in your app (or separate it)

with pixel-tracker there is no aggregated information, but there is a little more detailed information about each request

You can do something like this with mixpanel if you don't want to setup a whole server. I wrote about it.

http://alexkessinger.net/2010/12/17/doing-a-pixel-ping-witho...

I've never heard of pixel-tracking before. But Will wont it break if the user caches it? Then It wont be read again and thus not count the view.

Or?

the pixel doesn't get cached

Is that the reason why pixel tracking is used because other requests might be cached? When viewing a page, the browser client may just pull the page from cache and not make a request to the server? A 1x1 gif is small, static, and downloaded asynchronously and makes a negligible impact both client and server side.