Cites & Insights November 2005
Cites & Insights 5:12, November 2005, Walt Crawford's publication, was just released (and I'm remiss in that I haven't even managed yet to write about the previous issue). Besides e.g. good statistical skepticism of certain surveys, the meatiest part of the overall mix is:
* Net Media Perspective: Analogies, Gatekeepers and Blogging - some notes about net media and analogies, more comments on Civilities' "New Gatekeepers" series (and a related essay on citizen journalism), notes on seven other blogging essays and papers, and a few notes on Meredith Farkas' first-rate demographic survey of the biblioblogosphere.
This actually gives me an opportunity to write a little more deeply than my standard unheard messages about being unheard, and connect to some issues of search algorithms. The key passage:
[Jon] Garfunkel calls aggregatable declarations "crucial for markets and democracies" and says it's unfortunate that "so much of the communications essential to both democracy and markets escapes aggregation." He then goes on to note "practical deployments"--e.g. Google PageRank and Technorati rank. And here I see why I may be having so much trouble with Garfunkel's series--why I keep recommending it and talking about it, but disagree with so much of it. Garfunkel's looking for ways to establish significance. I'm more interested in discussion and complexity. I believe Garfunkel's looking for the kind of simple "good/bad" rating that aggregatable declarations lead to.
Well, one problem is precisely that "discussion and complexity" is often gatekept by "good/bad" rating, for the simple reason of time constraints: "Do I read this or not?". What informs the discussion? Hence the rise of gatekeepers. I'm not being extremely original there, but I suppose it's worth emphasizing that connection in this context. There's a whole structure underlying any system of discussion in the first place. It's more like whether you want to study the process of cooking, or review the quality of the meal itself. Note these are not completely disjoint, as in "This tastes bad (quality) because it's been burned (process).". So:
There may be structural problems that keep giving those who already have voices even more listeners, ... (Seth Finkelstein reacted to the "more flat society" possibility with some pessimism, mostly because it's such a difficult problem. "Nobody knows how to do good technology for non-hierarchical organizations...")
Right ("Nobody knows ..."). Here's where I connect to a simple worked example. Recently, Yahoo search started pointing to some blogs for "News". Now, I am arguably the world's expert on censorware - and if not, certainly up there. What I write is likely orders of magnitude more accurate than popular pundits. But my material won't appear in those search results (a yes/no decision). For the simple reasons that I don't have the voice that A-listers do (and, no, personal tone isn't the reason, that doesn't exclude the big blogfish). Which means the hierarchical organization just got a little stronger. No technology in widespread use measures my knowledge of censorware. I wish there was something that did, and someday in the far future, that may happen. But every time I have to go around pitching gatekeepers to be heard, the gap between what people would like, and what exists, is manifest to me.
By Seth Finkelstein | posted in cyberblather | on October 14, 2005 11:54 PM (Infothought permalink)
Seth,
Permit me to suggest that maybe you're using the wrong criteria to judge whether aggregation is successful. What I'm about to say might sound a bit circular, and perhaps it is in part, but it seems also the case. There are people whose business is, essentially, learning stuff. We call them researchers and tell the "common people" that they should pay for their food because they will come up with important stuff from time to time, cool stuff like iPods or better screen technology, but in reality most research has no bearing whatsoever in everyday life. In fact, the actual job of many researchers is, I postulate, to cultivate their mind, to learn stuff and create mental models and maps of that stuff. These people, whom I have called researchers but I could also call "experts", have the most "relevant" information on whichever topic you care to name, by certain metrics that you and I probably care about: accuracy, detail, consistency, completeness... But most people, when they look for non-specialized content (and most people reading blogs aren't looking for "expert" opinion in this meaning of expert) have different metrics: roughly we could speak of things like simplicity, conciseness, magnitude of "wow" per unit time spent reading. I'd argue that A-listers are, in a very narrow sense, "experts" in the creation of content addressed to such metrics, and "experts" in other matters are, for reasons obvious, not. Therefore it makes sense that an A-lister will be linked on censorware, and not you, under the premises above and under the premise that the aggregator is serving the user according to the user's wants.
There's a lot of critical analysis that could be made on this, I'm just leaving it here so you can think about it...
David: If I can paraphase, A-listers are experts at being popular, and data-mining is "successful" if it finds what's popular, regardless of whether or not it is right.
This can be a problem. Especially if data-mining is being put forth as the new new thing for civil society.
Right, your paraphrase is correct in essence which is why I think it's perhaps a bit of a circular idea. And I agree with you that for any kind of issue that actually matters, deciding on that basis would be stupid. So, I disagree about unsuccessful aggregation of content, but I agree about your scepticism that this is going to somehow open a new and shiny epoch of direct democracy(tm) or some such. An argument that the aggregators could make is that when people will start searching information to make decisions that matter and using it that way, the pattern will change. However I don't think that intentionality of people is easily aggregate()able at the moment, and until I see Google (or equivalent) finding out whether I want a detailed report to make my mind on a referendum or a sketchy sensationalistic post about it so I can show off some nice phrases to my friends on the same topic, I won't consider it feasible.