Microsoft’s “random” browser choice failure in requirements

I read a great breakdown from Rob Weir today on the issue of the “random” browser choice that Microsoft agreed to show users in Europe. For anyone who hasn’t been following this, the EU and Microsoft agreed to a set of rules and modifications to Windows that would give users a “fair” opportunity to choose a browser for the web. Or conversely, a set of rules that would allow browser publishers a fair chance to compete for market share on Windows.

The reality is that if you visit http://www.browserchoice.eu/, it is far from fair if you define “fair” as “every browser has an equal chance of appearing in the first slot, and an equal chance of appearing the second slot, and so on.” In other words, by any reasonable definition of fair, given the context of the case to begin with, what the EU probably wanted is a uniform distribution of the world’s top 5 browsers across the five columns on that page.

What they got was something totally different. And arguably, the IE team may have gotten the worst placement of all.

Now, part of it is that Microsoft’s programmers picked an awful algorithm. First of all, let me say that Microsoft employs some of the most brilliant minds in software that exist. Some (ok, not all) of the research and products that come out of Redmond are brilliant. But, unfortunately, when you’re a really huge company, not all of your programming teams will meet those highest standards. And even the best programmers can be sloppy when on a deadline. (although in this particular case, I don’t think they can claim the “we were in a rush” defense!)

Maybe Microsoft outsourced this particular bit of work. In any case, someone needs to be called into the office on Monday.

But what about the role of the EU’s negotiators and policy-makers in this? Here is the agreement, and the relevant section:

(13) The Choice Screen will prominently display the Final Releases of the five highest ranked web browsers based on usage share in the EEA (i.e. only these browsers will be immediately visible without requiring any user action under typical user settings). These five web browsers will be displayed in random order each time the Choice Screen is presented. The remaining seven browsers will be displayed if the user scrolls sideways and will also be displayed in random order.

There’s that word “random”. What they really meant to say is that the browsers should be uniformly distributed across the five possible positions. Now, again, I think any reasonable person would realize that “random order” in this case implies that the browsers should be uniformly distributed.

But if that is what was meant, then why not just say that? You can’t assume that the programmer sitting down to write this Javascript understands the legal and policy context. Somewhere in there is a good lesson for requirements analysts.