This facility allows you to register signatures of documents and check if documents with similar signatures have been registered in the past.

A signature is a set of keys and points that identify a text.

Keys: Two texts are considered to match if any keys in their signatures are identical. It is a good idea for keys to be self-identifying. For example, if you use the XYZ algorithm on a text and derive the key abcde, you might choose to use the key XYZ:abcde to identify that text.

Points: A point is a k-dimensional vector of numbers with a string identifier. If a text is registered with a point, then its nearest match for that identifier is that registered text whose point under the same identifier (if any) is closest in Euclidean k-space to the given text.

Name of text: Every registered text needs to be named. Names must be unique. Names identify the text so that if a new text matches an old one, you can figure out what the old text is. Some guidelines about names:

• Preserve the privacy of the original author. If you want to identify the author, you might want to encode that identification by using an identifier that can only be interpreted within your institution.
• You might want to include the institution, term, class name, instructor name, and assignment title; again, you may wish to encode that information.
• If the text comes from a web site, you might want to include its URL.

Submitter: Only the submitter who registers a signature may remove it. Submitter names are not revealed. You may use any string you like. I suggest you form it as a combination or your actual name and an encrypted password.

Author: Raphael Finkel (email, web).