My first experience with the Live Search OPML Generator -- Searching for feeds.

bobreb — Thu, 19 Oct 2006 07:06:40 GMT

This OPML Generator idea is clearly right up my alley. Kudos to Andyed for pulling it together.

I've been whispering about the relevance of social network placement in every Live(.com) ear that would listen for well over a year. I took my present position at Microsoft.com in part to deliver a service of that sort to IT professionals and developers. And in fact we are heading in that direction, but that's another story altogether.

I don't want to overstate the case with the OPML Generator. It's a Live.com search using the "Feed:" filter with some very nice features added that make it easy to select which feeds among the results returned should be saved as an OPML file.

Andy's timing was personally interesting. I had just delivered an OPML file to a colleague on the subjects of social software and web20 (that was strange experience). So it occurred to me to search the OPML Generator for social software and web 2.0 and compare the results with the collection I'd just delivered. There was only about a ten percent overlap.

All of my "top of the long tail curve" entries were represented, but as I should have expected, none of my farther out on the long tail entries made the cut. I decided this was not so good.

On the other hand there were a few feeds offered that I was not familiar with and about half of them turned out to be worth further attention. That was decidedly a good thing.

There were several feeds included in the OPML Generator generated list that are popular, well known to me, and intentionally not included in my list. I was undecided about that, but in the spirit of a marketplace of ideas determined it tended towards a good thing.

So, thought I, what's going on here. Why would the two lists be so different if both were supposed to accomplish the same thing? Especially if we consider that the ranking algorithms used by search relate to hits and links (in, out, and bidirectional). That sounds like exactly the sort of data set that would describe a social network.

I think the reason is related to how strange it was pulling my own recommended feeds list together in the first place. I realized, while studying my own feed reader and the many feeds I follow, that it really was my list. I'm not talking here about greedily holding on to my secret list. It's more like the list of feeds I follow are tuned to me. They fit the contours, so to speak, of my knowledge and experience. They've evolved over time in lock step with my own development.

Let me try to put it another way. I have feeds in my RSSBandit of bloggers whose opinions make me cringe. I almost never agree with anything they say. However, I came to understand some time ago that despite the aggravation I feel reading them, I almost always end up with a better perspective having subjected myself to them. If they knew about me, I like to think they'd feel the same way. I'm sure that's true as concerns the aggravation part.

Also, lots of my feeds are from "little guys" like myself; people that aren't linked to broadly; people that are not trying for blogger stardom, but are instead just chronicling their journey. They tend to have the greatest value to me.

As I've said elsewhere, top of the long curve bloggers generally prattle on about what I call water-cooler talk. I may have a social responsibility to remain conversant with it, but it has little or nothing to do with the things that occupy me most of the day.

So where does that leave the OPML Generator and feed search generally. It is at least a reasonable place to start. That's certainly true when you consider the alternatives. And in fact, I've decided to add a feed search to my reader just in case something interesting does show up there -- you never know. But this whole area is just begging for a better solution.

Process of Change : longtail

My first experience with the Live Search OPML Generator -- Searching for feeds.