Wikipedia:Anti-Vandal initiative

This page is a beta community interaction initiative page which is dealing with creating an algorithm which would have the capacity to circumvent vandalism to articles, in all its forms. I have spoken to people who actually create these algorithms and they first say that in order to have an accurate view of what vandalism is like, to try and define it into several different categories, so that we have an idea of what we're blocking.

Firstly, I think that the best option would be to have some form of algorithm that checks the grammar of the text that is being entered. First and foremost, because then it would be able to automatically check for bad grammar, fix what it can, and then mark up any other sections which are badly formatted. Once there is some form of algorithm for grammar, then we are able to combat vandalism most effectively.

Factors to take into consideration with the algorithm for vandalism

• Un radically changed words. For example, in an article I also write, words such as legumes, beans were switched with fags pork and beef. Now, how would a computer algorithm go about checking this difference which is easy to spot with the human eye?
• Create "Beginning phrases" that may be seen as "tags", such as "hi my name is" or "<person> is homosexual".
• Kill Edits that are exclusively in capital letters.
• Large changes in the character count for a given article.
• Words unrelated to the subject of the article (e.g. "cake n pie" in an article about handguns.)
• The percentage of edits made from Public IPs can be used establish a rank of the article's "attractiveness" to vandals. For example, a look at the virginity page shows half the edits from anonymous (young) users making bold statement about their friends past experiences.
• In parallel to the last point: IPs making a large proportion of their edits on "vandal attractive" pages loose "innocence credit". The two approaches combined would probably end-up making a form of "school IP detector"
• Add your ideas..