Based on feedback to the latest article and having thought about it some more I've gone and made some changes to the demo db.

The main issue with the approach was that it used a link. The problem here is that a web-facing application is prone to web crawlers adding their votes without realising what they're doing. If you're using it on an intranet then you're ok, otherwise it's not perfect.

To get round this I've applied the same technique to a button on a new form in that database. Getting carried away with myself I also went and made it in to a more robust and "real-world" system that could potentially be of actual use. It's not a simple survey site with multiple answers and coloured bar charts of the results.

As an example here's a test survey I created — What is the best web server environment? I wonder what the result to that will be?

You can download the updated code in the article. It's still far from what I think a proper survey template should be like. This is because I wanted to demonstrate the underlying principles more than I wanted to make it a plug-n-play template. That said, an upcoming requirement I have is to add surveys to a site (part of the reason I could afford the time to do this), so I'll be adding improving this database as I go along.