Mur J M, Viard D, Mathieu J L, Martin J
Med Inform (Lond). 1979 Jul-Sep;4(3):199-202. doi: 10.3109/14639237909010914.
Identifiers must be easy for access and discrimination. Usual identification by surname and christian name is convenient, for these two identifiers are almost always available. Their discrimination ability was studied in terms of the theory of information and the rate of homonymy. In the French language, the first five letters of the surname provide information equal to 12.11 bits and the rate of homonymy is about 0.659%. If one adds the first three letters of the first name, the gain in the quantity of information is 1.68 bits and the rate of homonymy becomes 0.087%. So the first five letters of a surname and the first three letters of a christian name ensure a relatively satisfactory identification and may constitute a significant way of reinforcing the discrimination power of another identification system.
标识符必须便于获取和区分。通常按姓氏和名字进行识别很方便，因为这两个标识符几乎总是可用的。根据信息理论和同音异义词发生率对它们的区分能力进行了研究。在法语中，姓氏的前五个字母提供的信息量等于12.11比特，同音异义词发生率约为0.659%。如果加上名字的前三个字母，信息量的增加为1.68比特，同音异义词发生率变为0.087%。因此，姓氏的前五个字母和名字的前三个字母可确保相对令人满意的识别，并且可能构成增强另一个识别系统区分能力的重要方式。
