The watermarking of images and documents is a common way to retain a copyright notice even if the file is copied. Indeed, a watermarked image with the website or name of the creator is actually an advertisement, even if unwanted. But watermarking text is notoriously difficult.
Watermarking Text
The text on a webpage is very different to the same text in a document format such as Word. Try opening a Word document in Notepad and it looks like gibberish. The text has been encrypted and adding a watermark to such documents to protect the author's copyright is straightforward to do. The same article on a website is not encrypted in any way and is best referred to as plain text.
Character Encoding of Webpages
Text on a website is encoded, but this is different. Character encoding is a protocol that allows the use of character sets beyond the usual Roman letters so that webpages can be created and read in a variety of languages, even languages with complex ideograms such as Chinese and Japanese. On Firefox one can see this in the menu item View>Character Encoding. Every webpage, even in English, has to define which character encoding method it uses – if it fails to do so then the default browser setting will be used.
Watermarking With Text Encoding
This has led some attempts to create watermarks by using invisible letters that are encoded into the document but are not visible in a browser. However, such methods can be seen if one looks at the plain text, for example by clicking View Source on a webpage. They will work if a plagiarist blindly copies and pastes the text but can easily be spotted, especially if the text is fairly short. It also requires some knowledge of coding or the use of a script to automate the process. However, most writers want to concentrate on their writing and have a method that is both simple to implement as well as simple to check if and when an article has been copied without permission.
Watermarking With Non-Standard Characters
Look at this phrase: it's one simple method. It looks perfectly ordinary, can be copied and pasted and still look perfectly ordinary, but some of the characters used are from the extended character set, such as Russian characters. There are many letters that look identical to the Latin ones but are actually from other languages and will therefore have a different code. They look the same to a human but will appear as different to a search engine.
Using Google Alerts to Track Plagiarism
The would-be plagiarist notices nothing out of the ordinary but you can now use Google Alerts to warn you if this phrase is ever used again. The advantage of this method over simply using the Google Alerts function for every article is that you can generate maybe half a dozen key phrases using non-standard characters and not worry about adding a new Alert phrase every time you post a new piece of writing. The key phrases can also be relatively short and common instead of looking for long unique phrases for each article. This obviously won't stop the plagiarist but it will alert you to when it happens.
See also:
Using Google Alerts to Detect Plagiarism
