After two and a half years of a fascinating, challenging, and wild ride in the Windows Phone division, I am moving this coming week to the Developer Division where I will be Senior Program Manager in Visual Studio overseeing a project that spans the Visual Studio marketing and developer websites.
This will be the fourth large website project I’ve done for Microsoft, the first three were combining the Windows XP, Windows Vista, and Windows 7 TechCenters, launching Microsoft Answers, their first forum-based consumer customer support site, and launching App Hub, the Windows Phone developer website, which combined the XNA Creator’s Club and the previous Windows Phone developer site.
The common threads across all of these sites that I drove was re-factored information architectures, large-volume content presentation, and leading cross-divisional working teams. I’ve learned quite a bit over the past six years at Microsoft about these areas that I’ll certainly bring to bear in my new role.
On a different note, most recently, I’ve spent the past few months analyzing quite a bit of data around the API Reference portions of the MSDN Library and have been evangelizing my results to documentation teams across Microsoft. I’m cautiously optimistic that a couple of teams have taken the data to heart, as I know that some work has begun to address some of the larger pain points. When this work by many, many people eventually comes to fruition, it should dramatically increase MSDN Library search engine relevance, hopefully making problems like this and this much less severe, and make the treasure hunt for API information less frustrating.
In a nutshell, structural artifacts of the documentation process creates web pages that are similar in content. Search engines use a process called shingling to de-duplicate and willow results. In cases where they look across large, structured documentation sets, you may not ever get search results for specific pages because they look too similar to other pages. (Examples A, B, and C.) Mark Manasse of Microsoft Research was kind enough to give me some of his time in December to explain this in more detail and give me some great ideas to pursue to understand the scope of the problem and ways to solve it.
I wish I could say more now, but changing page patterns for millions of web pages takes time. I’ll keep you posted. 🙂