Ah, but how far can they go? The limit was tested last week when new Post ombudswoman Deborah Howell tried to correct her assertion that lobbyist Jack Abramoff had also given money to Democrats. That post was attacked by an army of liberals whipped up by the group Media Matters and other liberal bloggers. According to washingtonpost.com editor Jim Brady, two staffers couldn't handle the onslaught of comments, many of which were inappropriate for a public forum.

The comments were turned off indefinitely for the Post.Blog, though the other Post blogs still have comments. Shutting the comments set off another firestorm as media watchers complained that the Post had overreacted and was simply deleting comments it didn't like. Brady denies that this happened.

"What we're not willing to do is allow the comments area to turn into a place where it's OK to unleash vicious, name-calling attacks on anyone, whether they are Post reporters, public figures or other commenters," Brady wrote. "And that's exactly what was happening."

(You can read a whole roundtable discussion on the topic here.)

The whole mess has made me think hard about my own nascent move into blogging. The folks at PBS.org and I have gone back and forth on blog comments policies. At first, I was going to check each and every post before it went online. Then we decided to allow people to post directly after giving their name and email address (without verifying either). We are still talking about other methods for moderating the comments, including obscenity-blocking software filters or spam-catching filters.

So far, we haven't had a bad start with comments here, but just wait till the political season kicks into high gear in the fall... You never know what's to come.

Let's lay out the various scenarios for monitoring online forums, blog comments and other public spaces on the Internet. Then we can think about just what we might do here at MediaShift, and what others like washingtonpost.com might consider in the future.

Scenario 1: No moderation
Pros: Everyone gets to participate with no censorship.
Cons: Give it enough time, and it will devolve into a messy, name-calling morass.
Example space: Many usenet discussion groups, and other non-commercial discussion boards online.

Scenario 2: Self-policing community
Pros: No time outlay for site owner; community members feel empowered.
Cons: Requires technology filter so people can flag and rate other people's comments.
Example space: Slashdot, MetaFilter group blogs.

Scenario 3: Moderate filtering
Pros: Keeps some bad eggs out by using technological filters and human watchers.
Cons: Doesn't keep all of them out, and might seem random in censorship.
Example space: Most news site blogs and other high-trafficked commercial forums.

Scenario 4: Heavy technical filtering
Pros: Forcing user registration or paid subscription will keep commentary to a select few people who care.
Cons: Doesn't allow people on the fence to join in easily.
Example space: Some newspaper sites such as NYTimes.com and narrow-focused sites.

Scenario 5: Heavy human filtering
Pros: Someone looks over every post to make sure it is on-topic and not a personal attack.
Cons: Might slow down online conversation; could cost site owner more money and time.
Example space: Many citizen journalism efforts such as at the BBC and MSNBC websites.

If there would be one great use for artificial intelligence, it would be to create a monitoring robot that could spot the flamers, trolls and other cretins who disrupt online forums. But in the interim, we'll just have to consider our audience, consider what ground rules to set, and listen to your input on how to make it work without making you jump through too many hoops. Just like every other online forum, we'll walk that fine line between giving you power and hoping it doesn't blow up in our faces.