Tagging, Search, and Re-Search

Filed under: sphere, tagging

The SupraSphere tagging engine is one of the most unique and innovative features of the system. We have taken a very different approach to tagging, almost redefining it to the point that we could reasonably call it something else!

In any case, our tagging engine has numerous different and highly compelling attributes:

1. When tagging an item, you can use multiple words in a single tag
2. When an item is tagged multiple times, it creates a relationship between the different tags
3. Tags are included as a part of search, so that even if a message, file, email, contact or other asset type do not have the keyword that is being searched, they will show up in the result set as a result of matching a tag
4. You can tag items in a search result with whatever keyword was searched for, making it very easy to associate new tags with assets just through normal usage
5. From a tag, you can get to all items associated with a tag very quickly, including other tags, making it much faster and more efficient than folders almost ensuring that you will never not be able to get back to your information
6. Tags also work as hyperlinks. If you are in a group sphere or different sphere other than your own personal sphere, you can get an auto-complete listing of all of your historical tags, at which point anyone else in the sphere will be able to see all items that you have associated with the tag, including other tags
7. In “Anonymous SupraSearch” (you can configure SS so that it exposes a search engine of all indexed bookmarks across all spheres), it will show all tags associated with any items that comes in the results page, so that you can click to find highly related items to whatever shows up in the results
8. There is something we call “Re-Search” that is a configurable option that will go into either your own personal tag repository (located in your personal sphere), or other users’ spheres, and highlight any word on the page that corresponds to a tag
9. Contact names are automatically treated as tags, so that you can search either one degree or multiple degrees away to see if you or anyone in your supra sphere knows anyone written about in an email or on a web page
10. You can match against a history of tags of everyone in a sphere. This is similar to re-search, but allows matching on an object basis.

These are just some of what the tagging engine can do. Tags are integrated deep inside and across the entire system, and we have many more plans for how we will use tags, such as for keyword alerting, monitoring RSS feeds, and for SPAM filtering. In many respects, the tagging engine makes our back-end architecture function very similar to a database filesystem. We would sincerely welcome researchers who are interested in experimenting with new concepts around tags!

The Last Mile of Search.

Filed under: search

“Oh, just Google it!” is the answer I often get when I can’t find something online. Sad to say, even with Google taking over the world, it can’t find everything. I could easily spend all day searching the wonders of the internet but where do I store what I find? What happens if I find something that may seem irrelevant, but I need it later on down the line? Sure there are web services that enable you to customize search (Rollyo and Google Custom Search come to mind) which are all “personalized” to some extent. I’m all for social bookmarking sites, using various add-ons to Firefox to quickly add to del.icio.us or Digg or what have you. But still there is something big lacking.

There is this term, “last mile”, which although has to do with actual physical connectivity, it also has to do with end users of the internet since it’s the part that affects individuals most directly. Even if you have no problem with Google owning your browsing history, bookmarks, and identity, there are still strong reasons why there should be a “last mile” of search. Without a high level of control over the search and browsing environment, allowing *bottom up innovation*, search will only get more personalized, it will never get personal.

And thankfully, that’s exactly what SupraSphere does: complete control over your searches.

I was talking to David today, about Dell of all things, and then I did a “SupraSearch” to find absolutely everything involving Dell that we had ever talked about for as long as I’ve been using SS. Here is what I thought was cool, the fact that I could choose what type of communication to search through. I set it to search for “dell” through all available message types, and up popped a dozen or so matches, including a few bookmarks and an email. Now I could have narrowed it down more, but since I’d never *really* used it, I just wanted to see what would come up.

If you take a look at the image to the left here, *points enthusiastically* we have the search box that comes up and you can see the options I have as a user to do the search for Dell. This way I can find the exact instance I want. If I just wanted to see the bookmarks, I could check just the bookmark option. Now, what happens with the results? A new tab is opened up revealing the results that have hyperlinks back to the original communication.

This will vastly improve web-browsing in general because of the easy searching of literally thousands (at least in my case, lol) of bookmarks, not to mention the logging of all messages in your various spheres. I happen to lose things easily, but this way we’ll never lose any communications!