Date: Wed, 7 Jan 1998 02:30:20 -0500 (EST) From: Gerald Oskoboiny To: hwg-ops@hwg.org Subject: Web-based co-moderation (was Re: Lists reminder (Was Re: Preliminary Data...)) In-Reply-To: <3.0.5.32.19980106193903.00932560@mail.idyllmtn.com> On Tue, 6 Jan 1998, Kynn Bartlett wrote: > Last of all, I'd like to point out that many people -- including > myself, the current LG management, and most (if not all) of the > Board consider the "open", unmoderated nature of the Guild's lists > to be a major plus. This is exactly why they're not moderators, > and never will be (at least if I have a voice in it) -- because > the unmoderated nature of the lists is a valuable thing, and > moderating it would do great harm. The List Guides are a com- > promise solution; not the ideal, but the best that has _ever_ been > suggested by _anyone_ so far. I absolutely agree with everything written above. This should be required reading for anyone suggesting changes to the LG program. The LG program is in place because nobody has proposed anything obviously better. Nobody has proposed anything obviously better because the problem we're trying to solve is a difficult one. Until someone proposes something better, all we can do is defend the existing system and allow it to continue to evolve. So, I have a specific proposal. I wrote this a few weeks ago, and I wasn't going to post it until I cleaned it up a bit and until I knew I'd have the time to defend it after posting it, but what the heck... As you probably know, occasionally we experience problems with our mailing lists, like off-topic posts, flamewars, and poor netiquette leading to frustration for many list subscribers. We have tried to solve these problems -- and largely succeeded, IMO -- with the List Guide program, but once in a while we still experience a few problems with this system. We shouldn't feel bad about not being able to solve these problems easily -- the entire net community has been experiencing these same problems for a number of years now, and nobody has really been able to solve them to everyone's satisfaction. The basic problem is that the Internet is changing from a society where everyone shared common goals of openness, cooperation, and trust, to a society, well, resembling the rest of the world. :( There are a number of different solutions being experimented with to fix these problems (for example, http://www.boutell.com/boutell/usenet.html ), but usually the problem comes down to this: the more restrictions you impose on members of the community, the less noise there is, but if you impose too many restictions, nobody wants to participate any more because they no longer have any freedom. So far the HWG has gone with a relatively "open" system, where anyone can post whatever they want, and if they post something incorrectly, they get admonished for doing so afterwards. This is good in that it allows free speech and open dialogue between list participants, but bad in that once someone has violated the rules, the damage is done -- their off-topic post has already gone to the whole list (usually over 2,000 people, for our lists.) An alternative to this open kind of system that has been suggested and rejected many times is "moderation", where all list posts are forwarded to a single person who reads and approves or rejects each one before it shows up on the list. Moderation usually results in high-quality lists, but the atmosphere on the list is fundamentally different -- it just doesn't feel as open or interactive as an unmoderated list. Some of the specific problems with traditional moderation are: 1. It is a large amount of work for a single person. 2. The amount of time between a message being sent to the list and showing up on the list is greatly increased, due to the extra intermediate step of requiring a human's approval. This delay is typically between 8-24 hours (more if the moderator gets sick or goes on vacation!) 3. The content of the list itself is largely influenced by the judgment of a single person, and after some time the entire list may take on a different character as a result. (and that character may or may not match one particular subscriber's ideal for the list.) 4. The moderator gets burned out. Fast. I think many of these problems would go away if we had some software that allowed the lists to be moderated by a panel of people rather than a single person, using a Web-based interface. I call this idea "web-based co-moderation". Here's how I think it should work: When someone sends mail to the list, it gets put in a queue of messages that need approval before they are distributed to the list. Each list can have a number of moderators that share the task of reading and approving messages. Also, the approval process can be based on a simple voting system, where each co-moderator has a vote on whether a message should be approved or rejected. Possibly, moderators with more experience could have a higher-weighted vote than others. This system would solve problems 1, 2, 3, and 4 above, viz: 1. It would not be too much work for a single person: each list could have up to 5-10 part-time moderators, and none of them would ever be obligated to check the queue within a certain time interval the way a traditional moderator works: if people get busy or go on vacation, it doesn't matter -- others will approve the messages for them. 2. The delay between the time a message is posted to the list and when it is actually distributed to the list is decreased, because with a large team of moderators chances are at least a few of them are online at any given time of the day (especially with volunteers distributed around the world.) 3. No single person has complete control over what gets posted to the list and what doesn't -- as long as a message gets a certain number of votes (say, 2 or 3), it gets sent to the list. 4. A single moderator doesn't get burned out because there is no single person with a large amount of direct responsibility or daily workload. If an individual co-moderator burns out, it doesn't impact the list because there are others to pitch in. The interface could offer a number of other improvements over the current system, too: software could be written to auto-detect common problems like netiquette violations (too much quoted text, MIME attachments, overly-long signatures, ...), providing the moderators with a single-click way to reject messages for common problems. A single click could also redirect the message to another more appropriate list. Here are some rough thoughts on how the moderation interface could work: ---------------------------------------------------------------------- Message#: 1234567 Subject: web-based co-moderation Status : pending, 1 vote needed From : Gerald Oskoboiny Suggested action: [REJECT] due to length Possible actions are: [APPROVE] [APPROVE], but redirect to: [hwg-html] (pick one) [REJECT] as not appropriate for any our lists [REJECT] due to discussion of prices [REJECT] due to posting copyrighted material: ask poster for evidence of the right to re-post this material [REJECT] due to netiquette violation: [over-quoting] (pick one) [REJECT] and give this person a virtual wedgie [REJECT] with comments: ____________________________________ ---------------------------------------------------------------------- There would also be an overview page that lists all the messages in the queue. Taking action on a message would bring up the next pending message, to allow for speedy moderation. If more than one moderator uses the interface at once and one moderator tries to approve a message after it has already been sent to the list, the software will just inform them, and not send a duplicate to the list. The software could also be enhanced to monitor the history associated with each e-mail address and be set to auto-approve any messages sent by someone who meets a certain criteria of quality (ratio of approved posts to rejected ones, maybe a decaying average or something.) We have the server resources necessary to support it already, and I could write the software myself in a day or two. For the moderators, we could use our current staff in the List Guide program (I think this type of system would take less human effort than the LG program, which would free up some of the excellent LG volunteers for other HWG projects.) We could probably just try it with one or two lists at first to see how well it seems to be working before implementing it with other lists. What do you think? I've been mulling this over for about a year now, and have been itching to write the software to support it. I think it would be really cool, and solve all the world^W HWG's problems. :) [Wow: actually, now that I check, it's been much more than a year: I first started thinking about this in June 1995! http://www.hwg.org/msgid/96Feb6.231444-0700_mst.138918-1+100@amisk.cs.ualberta.ca ] Gerald -- Gerald Oskoboiny (GB) Member, The HTML Writers Guild System Administrator, hwg.org http://www.hwg.org/