|Think about Loose Coupling|
Re^3: How should Perlmonks deal with Plagiarism?by Withigo (Friar)
|on Oct 09, 2006 at 06:30 UTC||Need Help??|
If you tell the PM-terruhrists what methods you are going to use to smoke them out, then they will know what to prepare for to beat you!
What's to stop a clever plagiarist from doing this:
1) set up fake accounts, scrape for and then repost questions with high XP, and then from other fake accounts, scrape the answers and reposts them
2) then from the account they actually want to accumulate high XP, send you and your cohorts a message pointing out the the obviously plagiarized posts
3) XP PROFIT!
But in all seriousness, perhaps any new posts should be scanned for similarity to previous posts. I recall a few months back seeing a snippet or a meditation about a writing a text plagiarism detection module, I think it was by Ovid?
You could also add something to the XP scoring so that posts more similar to other posts don't get as much XP. And then also take into account the number of replies, and the similarity compared to all of those replies. Genuinely interesting posts tend to have many unique replies.
Ultimately I think this is an unintended consequence of the recent XP scale changes. One might think it's a huge disincentive to cheaters and participants to make a saint require 10,000 XP and then adding a whole heirarchy of levels higher than saint with XP levels that are impossible to attain. But perhaps the opposite occured--since high XP is now such an impossible attainment, it became valuable enough to cheat for!