I read Cliff Figallo and Nancy Rhine's May 2001 article, "Tapping the Grapevine," with a mix of agreement and bemusement-and some historical argument. Professionally developed "content" has been around on the Internet since before the Web became prominent-but it's also true that "online conversation" has been with us since the early days of the Net. I find it refreshing that the dominant use of the Internet is still email: one-to-one and many-to-many communication continues to matter.

I don't believe in "the commercial Web" as a standalone entity. Commercial portions of the Web mix seamlessly with noncommercial portions; loc.gov may not get as many hits as yahoo.com, but it's a higher-quality source for those who use it. That's beside the point-which is that substituting user-generated "content" for what Figallo and Rhine call "manufactured content" may be a tricky road to continuity, profitability, and respectability.

Let's call "manufactured content" what it is: Professional writing that has (with luck) been professionally edited. If writing experience and editorial oversight don't give "manufactured content" an edge over user commentary, something is terribly wrong. Either freelance writers have gone bad, editors have given up, or readers have lost all sense of discrimination. I don't believe any of those to be entirely true.

Gaming the Ratings
There's nothing new about using aggregated user feedback as a ratings mechanism, and it's hardly unique to the Web. Condé Nast Traveler turns its reader surveys into several major articles each year. PC Magazine and Consumer Reports both base reliability rankings on reader surveys. On the other hand, these reader-based reports are based on large aggregations: Results won't be listed if there aren't enough reports to be statistically valid. Neither Consumer Reports nor PC Magazine bases actual evaluation on user feedback; both rely on expert testing.

Similar techniques can work on the Web as part of an overall ratings scheme. NetFlix offers an average "star" rating for each DVD based on user ratings-and makes those averages meaningful by noting the number of ratings for each DVD. NetFlix encourages viewers to rate DVDs by making it easy to do so, by offering a one-click-response "rate it now" email each time you return a DVD, and by offering custom recommendations that theoretically get better as you rate more DVDs.

NetFlix also shows one problem with user-generated content, as do many of the user reviews in Amazon. Textual reviews submitted by users can be useful-but they can also be peculiar, as when one viewer gave West Side Story one star because he or she was expecting an action picture, not all that singing and dancing. When one misguided jerk is mixed in with 5,000 others, it's irrelevant-but when it's the only user review in Amazon and the user is coherent enough to hide his or her ignorance or prejudice, the joke is on the reader.

So three reviews of an electronic toothbrush give the highest possible rating and say wonderful things. How many of those reviews came from employees of the toothbrush company? In Amazon, presumably, they can't all be from the same person using different registrations (or can they?), but many sites relying on user content have no such safeguards. Were there unfavorable knowledgeable reviews of the toothbrush that disappeared from Amazon? I've seen book reviews appear, disappear, and reappear with no particular pattern. I know from personal experience that Amazon sometimes suppresses authorial information; why should I assume that all user-generated content is treated fairly or equally?

More than one Web site offers reviews of cruise ships and aggregates numeric user ratings into overall scores. I've read enough of these reviews to be deeply skeptical of such scores. One review of one of my favorite ships (one that consistently rates among the top 10 ships in the industry) makes it sound like an aging rustbucket with mediocre food and lousy entertainment-while another review, from another passenger on the same cruise, makes the ship sound like heaven on earth. I've seen 150-square-foot cabins described as larger than average and spacious while 220-square-foot cabins are described as cramped.

Aggregate ratings can serve as useful information, but aggregate ratings should not substitute for professional writing. Amateur reviews can be lots of fun and might even be useful, but even if treated fairly by the sites, they too often tend toward rants and hobbyhorses. That doesn't make them worthless-but don't expect me to pay for that "content" or regard it as a reason to stick with your site.

Online Discussions & The Heat Death of Discourse
I was amused by Figallo and Rhine's characterization of Salon's Table Talk as "a salon for budding writers and populist commentators." There must be some of that, but I've encountered much more of the Freeper Phenomenon: Countless flames from extreme right-wingers with lots of time to troll the topics and good typing skills, aided by near-total disregard for logic, grammar, and spelling.

Perhaps there are specialized Table Talk topics that bore the crazies enough so they leave; I just haven't seen many of them. But then, I've stopped looking; who has the time? Anything remotely connected to politics or social issues seems to bring the usual suspects out of the woodwork. Blessed with multiple identities, an array of cute pseudonyms, and the charming (but infrequent) quirk of tracking down people who disagree with them and don't use pseudonyms and harassing them in real life, this coterie is another great weakness in "user-generated content."

At one point, I was moderately active in The Gate's conferences-but I fled from the barrage of vicious attacks and "you're another" responses. I've seen the same happen in most broad-based online fora. If Salon is becoming dependent on user-generated content, then Salon is failing. As it is, Salon's professional contributors include skilled ranters along with a few good writers-but David Horowitz and Joe Conason are nuanced and well-reasoned compared to the amateur posters.

Can conferences avoid the heat death of inflamed discussion? Maybe. Define the topic narrowly enough (and apolitically enough) and the extremists will go elsewhere-but, as ZDNet's old feedback mechanisms showed, there are extremists of all stripes. If other posters don't respond to the nutcases, they'll go away, eventually, but others take their places. Make the discussions technical enough (as on slashdot), and the readership may eliminate most political extremists-but not all (as on slashdot!).

Then there's moderating. It works, but there goes the free lunch. Active moderating-reviewing messages before they appear and rejecting those that are worthless or off-topic- slows the flow of discussion and can cause its own problems. Passive moderating-allowing automatic posting but kicking people off the system if they're out of line-works well enough in email lists, but poses another set of dangers.

Not Hopeless, But No Magic Bullet
I believe in lists, discussions, and other forms of user-generated content. I don't believe that such content can plausibly replace professional writing, at least not without substantial editorial oversight. As a consumer, I don't see freely generated content as part of "the business environment"- I see it as part of the Web that I pay for with my $19.95/month ISP payment.

You can enrich a professional site with amateur feedback in a number of ways. That's great. But if that's most of what's there, I'm gone. So, I think, is your business model-unless you're Topica or something similar.