Collective code ownership

Speaking of rotation … what happens when the same principle is applied to other types of teams? For example, software development teams.

You could argue that we use a rotation policy of sorts within the dev team at Trade Me in that we tend to mix up the projects a bit so that over time everybody works on different parts of the site.

This is a form of collective code ownership, which is not a new idea.

The main benefits are anybody is able to make changes to any part of the application, without fear of stepping on others toes; no individual becomes a bottleneck when changes are required; and the team is resilient to changes in roles and personal (this was also the justification used by Graham Henry for his rotation policy, pointing at the impact injuries to key players had in previous World Cup campaigns).

But what are the associated costs?

As Stefan Reitshamer points out, there is a pretty fine line between everybody owning the code and nobody owning the code. Instead of maximising flexibility and code quality as intended it becomes a tragedy of the commons.

I’m not sure there is a simple answer to this problem. However, unlike Messrs Henry and Bracewell, it’s nice to be able to work through these trade-offs without the media scrutinising every decision.

Trade Me browser stats for December

In December our three sites (Trade Me, FindSomeone, Old Friends) combined received just over 66% of all domestic page views recorded by Neilsen//NetRatings.

So, our server stats are probably the closet thing there is to a census of the technology kiwis are using to access websites.


Browser Market share Dec-06 +/- since July-06
IE6 70.3% -12.3%
IE7 12.2% +10.8%
Firefox 1.5 7.1% -0.3%
Firefox 2.0 3.8% +3.8%
Firefox 1.0 1.9% -1.2%
Safari 1.6% +0.5%
IE5.x 1.4% -0.7%
All others 1.7% -0.6%

There has been a lot of change in the last couple of months following the release of IE7 and Firefox 2.0, which together now account for over 15% of our visitors.

Despite all of the good press that Firefox gets within the web development community IE is still dominant with around 84% market share. IE7 has quickly jumped to 12%, no doubt thanks to Windows Update. It will be interesting to see how this tracks over the next few months once all of the users who will receive the update automatically are accounted for.

Screen resolutions

Screen res. Market share Dec-06 +/- since July-06
1024×768 54.1% -2.0%
800×600 15.2% -4.7%
1280×1024 13.2% +1.2%
1280×800 6.6% +2.2%
1152×864 2.9% +0.1%
All others 8.0% +3.2%

There is a slow but clear shift towards larger monitors. At the top end things fragment quite a bit, with various different sizes to consider. Over 30% of our visitors are now using 1280x or bigger. But that still leaves a majority using smaller resolutions. It’s depressing to think that many of these people probably have a monitor capable of a higher resolution, but are unable (?) to change the setting.

How does your setup compare?
Personally I prefer Firefox. I switched when Firefox 1.0 was released and haven’t been tempted back. I use a 19″ monitor which runs at 1280×1024.

Which raises some interesting questions for web designers, developers and testers:

If you develop using Firefox do you really have your users in mind? At Trade Me our test team all use IE6 as their primary browser, for reasons that should be obvious looking at the table above.

Have you already abandoned IE5.x users? This is still a big audience. 1.4% of Trade Me users represents around 40,000 unique visitors each month.

How does your site look at 800×600? These stats are a reflection of our audience, but they’re also a reflection of our site. The new Trade Me design, launched in November, targets users with a resolution of 1024×768 or larger, but we’re also careful to ensure that the site still works for users at 800×600. On the homepage, for example, the size of the category links is reduced and the navigation tags are repositioned below the logo and banner advert. If your stats show a lower percentage of users with small resolutions, why is that? If your site works poorly for these users they are unlikely to come back.

How do your stats compare to these?

Source: all of the numbers above come from Neilsen//NetRatings.