From: Geert E. <rou...@us...> - 2008-05-24 06:54:32
|
I can only guess, since I only read this. But the most logical would be that if you have every user on moderation queue and when an user submitted once a valid non spam comment, he gets a special flag so he doesn't need to be moderated the next time. But it can only work for registered users. The pros: saves time since you only need to moderate new users, not every new comment. Cons: user needs to be logged in. Now that I think of it, you could add an extra flag for users that are known for using offensive language. They can post their comment, it is visible, but also notifies an editor so he can check their comments right away. It would be a it like auto marking comments as offensive. Couldn't just a "confirmation e-mail" be dispatched with a random validation-url that the user must click to to "confirm his comment".... Could also be a solution. And more friendly then captcha's. I saw this method a few times too. I think if you allow guest comments, the only option is captcha's. On Sat, May 24, 2008 at 3:54 AM, Fernando Cassia <fc...@gm...> wrote: > On Fri, May 23, 2008 at 8:34 AM, Geert Eltink > <rou...@us...> wrote: >> >> Just ran into a page from wordpress with how they fight spam: >> >> - Akismet >> - Limit number of links in posts >> - Spam words >> - Comment blacklist (already in props) >> - Moderate all comments >> - Pre-approve comments from old commenters >> - Restrict comments to registered users >> - Inform users when comments are moderated by a (popup) message >> >> >> My favourites are a combo of >> - Moderate all comments >> - Pre-approve comments from old commenters >> - Inform users when comments are moderated by a (popup) message >> >> Or >> - Restrict comments to registered users only >> - Add an option 'report comment' so an editor will be notified and can >> modify or delete the comment > > > How does the "pre-approval" for old commenters work? I've never seen that > feature. > What prevents others from posing as legit commenters?. > > Couldn't just a "confirmation e-mail" be dispatched with a random > validation-url that the user must click to to "confirm his comment" serve > two purposes: > > 1. Prevent comments "identity theft" - ie someone else putting comments > under the name Geert Eltink in a story by Geert Eltink. > 2. Only allow comments from people who do not lie and provide a VALID email > address (checked every time, and with a random URL - if user doesn't click > to the random URL within 24 hours, it would be deleted from the system). > > Would therebe any drawback to this approach? > > FC |