I'd like to eliminate the current "Moderation Queue" entirely.
I'd like to replace it with a "Submission Queue".
The difference is that a submission queue is for ironing out the kinks in an article. Mostly structural issues -- spelling, grammar, broken links. Possibly some content issues -- there is a lot of, frankly, sophomoric crud spewed forth on K5 that's little more than uninformed opinion. Or maybe it is informed opinion, but the author has:
- Grossly mismatched a title to a subject.
- Failed to provide any external references.
- Failed to organize the article at all
- Gone hopelessly beyond his/her depth.
The submission queue is where stuff is submitted for a first, very rough cut, and slapped into technical shape. Really bad stuff -- spam, trolls, hopelessly lost articles -- are killed or remanded to diary. Everything else goes to section (possibly on the basis of what section the SubQue folks thought appropriate, possibly with a title SubQue folks thought appropriate). The only truly doomed content is spam and trolls. Everything else continues to exist and be accessible.
Once content hits the deck, it's still subject to moderation. Possibly also implicit measures such as how often it's read, how many comments are made, how highly rated these comments are. An article that generates a spew of flamebait doesn't merit front page placement. An article that generates a few, very highly regarded, comments, might.
The idea is to make the initial go/no-go decision quickly. Preferably within a few hours -- say anywhere from one to twelve. I'd prefer not to have anything hang in queue for more than a day. It's far less important to make critical decisions about quality than to promote a fair bit of reasonably good stuff.
Once on deck (in section), moderation and activity could further promote or demote an article. I'd like to see it be possible for an older article to be revived if it gets promoted high enough.
Periodically, highly-rated section content gets promoted to front page. Maybe once a day, maybe three times. Maybe once every other day. Scoop site manager gets to choose. This would include all sections, including best picks from diaries (it's really hard to find the good stuff in diaries right now). There would also be tools to provide lists of current, highly-rated articles within each section, etc. Scoop's an immensely rich data source, I just wish it would use what it's got.
The implicit assumptions are:
- It doesn't take too many opinions to get a basic assessment (very rough cut) of quality.
- No simple, computationaly unintensive, predictive metric is going to do a good job of predicting story performance.
- A better assessment is to look at first-day's performance on deck. Note that this is less time than many submissions currently spend in queue.
- A sufficient number of honest users exist to make this work.
- Abuse significant enough to effect the system is significant enough to be detected.
- Scheduling and other needs are handled outside the submission queue.
Karsten M. Self
SCO -- backgrounder on Caldera/SCO vs IBM
Support the EFF!!
There is no K5 cabal.
[ Parent ]