A Tale Of Two Threads

Here's a little compare-and-contrast exercise that really showcases what's great and what's awful about the web right now. On the one hand, check out this wonderful 100-post-long thread at Kottke.org. Jason launches the discussion with a fascinating series of questions about bilingual conversations, where speakers switch back and forth between different languages, often in mid-sentence. (Apparently, this is called codeswitching.) And then you have a discussion thread with a signal-to-noise ratio close to infinity, filled with personal anecdotes, speculation, further reading, scientific explanations -- contributed by mostly complete strangers from all over the world, chiming in because Jason asked some great initial questions. It's as rich as conversation as you'll find anywhere, and it's the kind of conversation that would have been for all intents and purposes impossible ten years ago.

On the other hand, you have the people posting junk comments in some of my old threads here, simply to boost the Google PageRank for their sites. It's not even pure spam -- it's meta-meta-spam: making a commercial pitch by making it more likely for people to find your site on Google by making it more likely that Google will find your find site by linking to it from other people's sites, thus boosting PageRank. Uggh.

By the way, I said all those nice things about Jason so he'd put me back on his blogroll. It really hurt my PageRank when he took me off...

Comments

[The Leach Boys]

Leach Blog September 15

Leach Blogs are all the rage. Everywhere you go you see them. Andrew Sullivan has done a few. Lawrence Lessig's blog got hit. There's been a huge debate about how often one should leach, is this a feminist thing, do leaches constitute intellectual property (who owns the leach), can leach blogs be used successfuly in English classes?

So much to think about.

[Fabio - from Brazil]

I've just finished you second (and great) book, and I really get interested in the intelligent agents stuff. Did you write something about it this year?

Fabio

[Sebastian Meyer]

I've had one of those comments too and could not figure out what they even intended. This explains it pretty well though.

And no, this is not intended to be one of those. ;-)

[Stephen Bronstein]

As Clay Shirky recently pointed out, Wikis, which are generally thought of as the most open and inclusive type of web site (since anyone can make any change), are only successful because of the rollback function, which allows the 'owners' to quickly rollback any changes that were not made in the 'spirit' of the Wiki.

Similarly, Mark Pilgrim has an extensive comments policy for his weblog, which is effective at eliminating the types of free-rider comments that have been a problem (or at least an annoyance to Steven) here. Eventually we'll have matrixed Whuffie scores to help us each personally filter content. Until then, pruning by management seems likely to be the most successful. Perhaps coupled with some sort of bayesian filtering (an MT plugin perhaps?) - bayesian filters seem to be licking the email spam problem very effectively, why not comment spam as well?

[reed]

Stephen Bronstein wrote:

bayesian filters seem to be licking the email spam problem very effectively, why not comment spam as well?

The filters in Apple's Mail.app program are pretty good: they used to catch about half of my spam. (I get lots). I am guessing that they are Bayesian.

spamcop.net is better. It blocks at least 98% of my spam without ever reaching my mailbox. (Spamcop works IIRC by running a regular DNS server. When people report spam to them they put the originating address in the server; I configured our mail server (exim) to check spamcop for each incoming mail.)

Intelligent filtering is good, but people collaborating (just by sending in a couple spams to spamcop) on the net is far more effective.

[Jeff]

It may be a Google PageRank thing, but a few of us MT users have gotten spam comments with links to "body-enlargement" and more benign sites.

When I did a lookup of the IP addresses, they were all coming from servers in China (and the email address attached to the comment was bogus). After doing the lookup I banned not only that IP address but a range of IPs from that host (the IP lookup will tell you the entire range that host uses).

[henrycopeland]

Unfortunately, if you are in Jason's blogroll javascript, I'm not sure you'll get the PR juice. :)