An example might be the following: "Romeo, wherefore art thou Romo?" from the classic Shakespeare play. Suppose I changed it slightly to "Roberto, why are you Roberto?"
Is there a computer program or algorithm that would detect the fact that I basically used synonyms for each word in the original, and thereby produced a high degree of correlation with the original? And would it be able to measure the "degree" of plagiarism?
On my last university course the department insisted everything was submitted to a system called "TurnItIn" (http://turnitin.com/en_us/home) but there are loads of others.
TurnItIn gave a load of different metrics for direct copies of other material, rephrasing and the like. Only problem I found was it often got confused with citations and references.
At my college they use this software to determine the authenticity of a thesis when submitted.
These softwares work by comparing strings of n words in your work to strings of n words in works in their database. Last time I checked (literally), they did not account for synonyms (which would be pretty hard to do), so they incite students to paraphrase (gratuitously) rather than just copy. (Why this is of educational value is beyond me.)
