Microsoft Research: Cutting-edge ideas for the cloud, trapped in a desktop mindset?

I have a new obsession (more on that below) and have been scouring the Web to see who else is already working on it (there’s always someone). The bad news is, in terms of running code, versions of the idea have been “in market” for years. The good news is that these products were developed by Microsoft Research, which means they’re unlikely ever to get implemented in the way I think they should.

My mental shortcode for the topic is “personal search”, summarized as follows:

Where’s the Pain?

  • As consumers and businesspeople shift their activity to Web-based platforms, they are losing control of their personal content archives (emails, files, contacts, etc.) due to fragmentation across siloed service platforms (e.g., Gmail, Flickr, Facebook, LinkedIn, Zoho, etc.)
  • As individual digital media creation expands and fragments (email, blog entries, digital photos, documents, bookmarks, etc.), it becomes harder to locate and retrieve individual items (or related collections of content) when required

Where’s the Opportunity?

  • Consolidated personal media storage and/or indexing in the cloud
  • Multi-modal search and retrieval from this personal content store (e.g., by date, source, file type, keyword / concept, etc.)

(I didn’t realize it at first, but as the idea has crystallized in my head I recognize elements of what Brad Feld has been calling “glue” in this idea, so thanks to Brad for planting the seed).

As I started digging into this idea, I found that all the best work (at least that I could find) on the topic had been done by Microsoft Research. In addition to the better-known MyLifeBits project, two related efforts – Stuff I’ve Seen (an early prototype of Windows Desktop Search), and a follow-on effort called PHLAT – are actually much closer to the idea in my head.

Being from Microsoft, both of these ideas are almost entirely desktop-centric (as opposed to the distributed and cloud-based implementation I’m envisioning), but the value prop is very similar:

“SIS is a prototype tool that makes it easy for you to find information you’ve seen before, whether it came as email, attachments, files, web pages, appointments, tablet journal entries, etc. We do this by providing a single unified stored of different sources and providing an interface with quick sorting, filtering, previews and thumbnails” – Description from the SIS page on

That’s it exactly, but the use case I want to support (and the resulting datastore) is entirely device-independent. All the data I want to aggregate and dig through lives in the cloud, including files I create in client applications but then pass to the cloud for archiving and ease of access. It’s not about sync (which is where Live Mesh seems to be focused), the cloud is the device.

I know of a few startups that are touching various bits of this elephant, but nobody who’s biting off the whole animal. Seen anything like this out there? Let me know.