Googlespeed Ahead

A short while ago Marc Orchant wrote a great post on Lookout, the Outlook search enhancement tool. I just, tonight, got around to actually installing it and completing the first indexing and oh man it's awesome.

I wanted to locate all postings related to "corporate transparency" which I first heard mentioned on the Scobleizer the other day. (I use NewsGator because it unifies my personal data access tool and allows me to read the entire blogroll on my PDA.) I just typed "corporate transparency" in the search box and it went to work. Normally I wouldn't have even searched for this because of the time it would take to locate; I would have just hunted around folders until I locked in on the correct posting. This time, however, in a mere 1.45 seconds I was staring at exactly the results for which I had been looking. I have to say, there's a certain feeling of power and productivity associated with searching your own information at Googlespeed.

Now I'd say the same thing applies to searching my hard drive as well. It takes an exorbitant amount of time to actually load the Windows search dialog and after the load completes, the time required to effectively search a catalog renders the search nearly unusable. Well, maybe it's usable, but it's certainly not running at Googlespeed.

Imagine the power you'd have if you could, given a certain subject you seek to research:

1. Perform a Lookout search on Outlook folders.
2. Perform an effective search of the entire file system.
3. Perform a Google search on the topic.
4. Union the three result sets.
5. Present a comprehensive result set, tailored to your search, comprising your personal information, your personal documents, and a cross-section of global intelligence procured from the web.

Wait just a second, that's WinFS. When Longhorn does ship, each machine will have that exact ability (and the ability to do it at Googlespeed!). Imagine the power and productivity boon garnished by such empowerment.

And so, in short, anyone looking to get a feel for how powerful WinFS will be in effecting our productivity need look no further than Lookout for a (very) preliminary and scaled down preview. Lookout certainly captures a small portion of the "wow!" factor of WinFS.

A glossary:

WinFS: (from Chris Sells and the Longhorn Developer Center) the code name for the next generation storage platform in Windows "Longhorn." Taking advantage of database technologies, Microsoft is advancing the file system into an integrated store for file data, relational data, and XML data. Windows users will have intuitive new ways to find, relate, and act on their information, regardless of what application creates the data. Also, "WinFS" will have built-in support for multi-master data synchronization across other Longhorn machines and other data sources. The platform supports rich managed Longhorn APIs as well as Win32 APIs.

Googlespeed: Searching a vast array (could be >1B items) at near-impossibly short timeframes. Otherwise known as the holy grail in most search markets.

Comments

[Kurt Greiner]

Try x1. It allows a similar experience today, in Outlook and also the flie system. Been using it for about 6 months and could'nt be happier.

www.x1.com

Kurt

[Paschal L]

The problem with winFs as I mentioned before with one post, is that you have to enter metadata to identify your own data. So that's mean entering data to find your data. I am not sure this is the right approach. It surely going to take a lot of time to index every single item from an hard drive and Bill Gates talk about 1 Terabytes for 2007. After all Google works pretty well without asking me to index my pages. I think this is a more clever way, using the computer for what it meant to be, an information retriever.

[Jeff Maurone]

An interesting point indeed, Paschal. However, it's not completely accurate.

Yes, there are some binary types you'll have to enter data to describe. For instance, your photos. How else can your machine effectively identify the content in a photo without some human input? Google Image Search doesn't do photo recognition, it looks for contextual information surrounding the HTML tags used to display the photo. Unfortunately such ascii content is not available in, for instance, your My Pictures folder. That being said, there are very "usability-conscious" ways to store this information about photos. For instance, right-click on a photo and select a contact to which to link it.

However, do you really think we'll have to enter metadata about any Office documents? Of course not, there will be translation and interface components used to effectively extract such metadata and store it in WinFS. I'd argue that most major-market applications would follow suit and provide WinFS extraction interfaces for their document formats.

[GoogleSpeed.com]

Why not try finding stuff faster on google

http://www.googlespeed.com