Personal Search Synergies

After the excitement of the last week, I’m finally catching up on work and subsequently blogging.

In particular last night I had a few minutes to check out one of the latest Desktop/Personal Search applications.

Specifically, Lookout Soft’s email search add-on for Outlook, which seems like a great tool.

In limited tests I found Lookout’s search accurate and fast (once the initial indexing was completed). In general I think Lookout and similar products such as X1 are immensely useful.

However, in my experiences the indexing functions of these personal search tools always seem to annoy me with larger corpuses — such as my fat 800 MB PST file — even on a relatively fast system.

So I end up uninstalling these tools because they simply get in the way more than they facilitate.

It seems I’m not alone with that opinion as John Battelle mentioned last week:

“Desktop search (ie searching your own hard drive) is one of those things that seems to have gotten worse in the past ten years (why Yahoo, MSFT or Google don’t do it is a mystery, imagine the goodwill…).”

As well what Philipp Lenssen said in reference to John’s post:

“We wonder why Google takes below a single second to find something in billions of pages (and do some clever ranking at the same time), whereas Windows takes anything from minutes to hours – for a small fraction of documents.”

In my opinion, I’d like to see more Personal Search applications borrow from architectures such as P2P, Social Networking and Grid computing.

For example, I imagine a cross between Groove, Eurekster and United Devices.

Specifically, I would take the secure networking and synchronous file updates from Groove and toss out its thick client.

Push indexing, data mining and analytics to a social or peer network via United Devices toolkit and wrap the entire package in a web service-based API that can be easily embedded into productivity applications such as those found in MS Office, Open Office and even Mozilla.

Indeed, I’m glossing over the details, but I’m sure something like this is either already out there or about to be released. Perhaps it’s where MSFT is going with Longhorn. I’m not entirely sure, but I am sure that there are obvious synergies with these technologies that I’ve yet to see completely tapped.