The BBC and Top Gear do, every year, a car survey. The 2005 Top Gear survey is pretty useful if you’re looking to buy a car, which we are. However, it’s organised appallingly; you have to walk through all 15 pages of it looking for the car you’re interested in. So, the Kryogenix International House of Work-Avoidance presents Top Gear 2005 Survey: easy access edition. All the cars are listed, by manufacturer, and you can jump to the page with them on. For those of you with a coding bent, you can see how all the work is done with two files, build.sh and parse.py. Essentially, it grabs the HTML for the survey, throws it through HTML Tidy to make it XHTML, and then parses it with Python’s minidom. Nasty screen-scraping, it is true, but it works. It extracts all the information from the survey, so you could use it to put together a more detailed summary with all the actual data in it (and not link to the real survey at all), but I think that’s a bad idea because the survey’s great and the BBC should get the credit for it. However, if anyone from the survey reads this, can I ask politely that next year (a) the HTML for the survey is a bit more structured and (b) you make car names individually linkable, so I can link to each car rather than just to the page it’s on? Cheers.