ASalina opened this issue on Dec 25, 2002 ยท 51 posts
CyberStretch posted Mon, 30 December 2002 at 9:48 AM
"but what images/posts would be 'flagged'?" Images would be the toughest, since I presume there is no scanning/pattern matching software that would be accurate/affordable/available enough to use. Therefore, to begin with, you could flag all existing images as being "ok" (or whatever variation) and any new images as needing to be reviewed. Posts could be flagged based off keywords, such as vulgarities, and specific terms used in the TOS that might imply a violation or at least merit some further investigation. For example: "fuck you" (case insensitive) could be one potential for the keyword list; since "fuck you" hardly ever describes anything constructive. However, "fuck", by itself, may be flagged with a lower priority since some uses of the word may be acceptable; trying to focus more on finding TOS violations vs acceptable uses of common words/phrases. Databases/scripts can also use pattern matching, so you could add "f?ck" and "fuck%" (or whatever the specific database/script uses) to limit the number of entries in the keyword list and avoid having to put in "fuck", "fucks", "fucker", fucking", etc. It could be coded to check the posts prior to posting, after posting, or when the post is archived, etc; the decision would have to be made which would be more fruitful and less resource-intensive or disruptive to the service. Generally, I would have the script run during a lull in activity; which could also be scripted (ie, IF MembersOnline <= 500 THEN run, ELSE wait). I would suggest categorizing and prioritizing the reports based upon what R'osity staff feels are more serious offenses. Ie, trolling may be more serious than advertising in a forum, so you would want to tend to the trolls first (higher priority); or have someone working on trolls and another working on advertising. Being mods/admins, I am sure that there are other such keywords or catch phrases that could be used that you have come across during your tenure. Allowing mods/admins to add to the list of keywords, it would be possible to expand the search generating closer hits. You could even add in a keyword/phrase for members to post that would trigger mod action if found. Or add a checkbox to the forum threads in which members could alert mods/admins by flagging the posts vs having to IM them. Although "member flagging" has the potential for absue and misuse, it could be a TOS violation with ramifications if used improperly; like yelling "fire" in a theater or "bomb" on a plane. If the checkbox idea is implemented, you could prioritize by the number of participating members who have "flagged" the thread, meaning that more members are concerned and request mod/admin intervention. I would suggest having the "flag" visible only in the mod/admin views; that way, confidentiality would probably produce more honest feedback and less retaliation. You could further enhance this by having a dropdown of common violations vs the checkbox so the members would be able to classify the "flag" and make it easier on mods to determine what to action first. No system would be failsafe nor foolproof nor would they be likely to work out of the box. However, adding the functionality to expand would help a great deal. By having the database, and perhaps members, do some of the more menial and mundane tasks, it would free up human resources to be alerted to or go after more potential TOS violations. How is this performed at the present? Is there a series of steps mods/admins take that could potentially be "reduced" to code? Again, I would caution not to automate the disciplinary actions due to potential mishaps with the script (like the Freebie removal script not so long ago).