There's a lot of reasons to use, or not use, any number of different identifier schemes. Let's say, for example, you want to say a file is a Word document. You could use any number of the following: an OID of 1.2.840.113556.4.2, a MIME type application/msword, an MS XML Schema http://schemas.microsoft.com/office/word/2003/2/wordml, plain old .doc file extension. There are doubtlessly other style available; MIME types having quite a few variations.

One reason you'd want to apply some sort of identifier is to allow finding it at a later time. If you use an extremely simple identifier you often end up getting more results than desired. If you use to complex an identifier you end up not being able to cast a wide enough net to find a large enough set to be useful. That and if you don't use a commonly shared identifier you won't be able to cast your searches across other sets of data.

What am I suggesting here? That recognizing many different forms of identifers exist and that, often times, creating new ones isn't always the best course of action. Sometimes participating in one (or many) different identifier schemes helps your data be useful to much wider audiences. That hassle, of course, is you have to take the time to understand what's available and how to use them.