The ietf-xml-mime mailing list is discussing the use of MIME content-type identifiers and their use with fragments of documents rather than complete documents. XML has muddied the waters once again.

The discussion is intended to refine an Internet Draft explaining how to describe XML in contexts where content-type identification is critical. MIME content-types were originally developed for email, but are an important part of the Web's HTTP protocol and are used in many other situations, notably to identify style sheet types.

At the apparent conclusion of an earlier discussion, Murata Makoto started by describing likely agreement from prior discussions - that fragments don't have media types - and then charged forward with some cases that may not correspond to that agreement.

Gavin Thomas Nicol suggested that media types of fragments might inherit a type from their parent document, but Bert Bos showed how this inheritance can't be figured out by applications using the W3C's style sheet processing instruction. XSL transformations in particular may be embedded in other document types.

• An earlier xmlhack article describes the background to this discussion.