Well, my other somewhat-experimental site, Knowledge Cog, has now reached 734 posts. A few of these belong to me, but the majority of them are the works that other people have created — and I’ve used Knowledge Cog as a public medium in order for other people to see the stuff that I like to read.
On a related note to this, I’ve heard some talk lately about the potential for a passive medium like television and an active medium like the web to intersect and eventually merge into one medium. While I don’t doubt that this is a possibility, I don’t think we’re as far along in a combination of these two media as some might think.
The technology to combine these two media already exist — watching a television program on a computer has been possible for quite some time now. Where the combination begins to break apart centres more around user behaviour than the underlying technologies however. While a site like Knowledge Cog provides an aggregation of content from various sites, and provides a rudimentary way to push content to a user, the web today is still mostly a pull environment.
Most of the time, we go to Google, type in some keywords, and pull what we want from the static web. This is far different from our experience when we watch television — where our choice as to what we consume are far more limited. Services that record live television are beginning to bridge the gap between these two media, but this is still far from pulling the two together into one.
As a species, we are beginning to run into human media-consumption limitations where technology is of little help. We simply have no way of consuming a great of deal of content at once — we can really only suck in one (or possibly two with picture-in-picture) television channel at a time.
Similarly, we can only go and search for one thing at a time on the web. Our information-seeking behaviour becomes simply: 1. Type in keywords. 2. Click search. 3. Browse results. 4. Repeat. Where we have begun to make some progress with technologies like RSS is in the fact that we can ‘watch’ certain keywords, and get updated every time they occur. Or we can be alerted every time a particular site posts a new piece of content.
So what’s the next logical step? Some sort of technology that makes some sort of basic, rudimentary sense out of what’s out there, and crawls the web not just with brute indexing in mind, but with simple sensemaking as well. We could have a personalized agent, with a basic level of understanding of what we want and like running through the web all the time. Let’s call this personalized indexing. I’d be interested to know: how far do you think this type of technology is from being realized?
Like this post? Subscribe now to the full RSS feed.
June 27th, 2007 at 12:05 pm
An interesting idea Luke. I too have an experimental Tumblr blog similar to KnowledgeCog. I spend hours bookmarking in del.icio.us and Digging (pardon the pun). It’s definitely a “pull” scenario — similar to KnowledgeCog, Tumblr pushes my ‘pulled’ content to my blog.
What you are suggesting goes beyond personalized “pull” search. Other than RSS and Google’s rudimentary personalized search service, I’ve yet to find anything that does both “pull and push” for me “intelligently”. So yes, in my opinion, this technology is far from being realized.
June 27th, 2007 at 1:11 pm
Great point Angela. I’ve also toyed with Google’s personalized search, and found it didn’t do much to assist me with my searching. I think that to take search beyond its current stage however, the major search players will (and Google certainly attempts to make strides in this direction) have to think outside of the box a bit.
The ‘intelligence’ factor in search (and we could say, in software in general) has been a long-awaited goal that never quite seems to get realized. But in a sense, I doubt there will be one revolutionary company or product that completely changes search, instead it will more likely consist of small advances that set our expectations ever higher.
As I was reading your comment, it made me think of something as simple as Google Adsense — the very fact that Adsense can not provide ads, but that it can provide contextualized ads by using the same crawl that it’s using to feel Google. The key may just be in tapping into other ways to make use of that crawl, and perhaps making it more robust.
Thanks for your comment!
June 28th, 2007 at 3:09 am
Very interesting idea….definitely opens up a number of possibilities..
June 28th, 2007 at 8:10 am
I’ve also been playing with tumblr as an easy way to aggregate my blog, del.icio.us and Twitter posts but then, widgets and RSS push much of the same content onto my regular blog anyway.
As far as personalised search goes, I’ve been experimenting with Particls which, whilst still very much a beta product takes my Google feeds (via OPML) and key words and tries to deliver relevant content (kind of like an ‘intelligent’ BBC Ticker). It works, more or less, and like a spam filter (or every other personalised search) learns through tracking what you click through to. On the other hand, it still feels very…obtrusive somehow.