This topic is discussed in Naming standards: file names; unique Ids. The focus there is on naming of files and on unique identifiers. Here we consider those matters briefly, but also discuss URNs and OAI naming.
At Virginia Tech, naming of the parts of an ETD is left to authors. Many have a single document, called etd.pdf, but there are many other selections. The entire ETD is referred to by way of a summary page. This has an ID like:
etd-12598-10640 or (in earlier cases) etd-5941513972900
These two correspond to full URLs, respectively:
Other but similar schemes are in use elsewhere - recall Naming standards: file names; unique Ids.
The key point is to have an ID for each work, and a way to resolve that to the actual document. In some cases, a URN is given, such as a handle or PURL, which in turn is resolved to a URL. (See Identifying: URN, PURL, DOI) That allows users to remember the URN, and be assured that over time, as documents are moved about, there still be automatic linking from the URN to the designated document.
IDs as discussed above also are important with regard to OAI. Assuming that, as is required, each archive has a unique identifier, then as long as each record in the archive, as required, has a unique internal ID, then the combination of the two will uniquely determine a record in an archive.
Next Section: Encryption; Watermarking