On 10/26/05, James M Snell <[EMAIL PROTECTED]> wrote: > David Powell wrote: > > >So I think it only really makes sense to do either: > > > > i) The spec doesn't mention MIME types (it just references XOXO). The > > assumption is that an HTML type is used, and nobody expects MIME > > dispatch to work. If we aren't bothered about MIME dispatch then this > > is fine. > > > >ii) The spec names a specific MIME type for the introspection > > document, the expectation is that all documents will have that type, > > and clients depending on MIME dispatch can expect to work. But, > > the XOXO document will probably not be browsable because it will > > lack an HTML MIME type. > > > > > > > First, let's ignore the references to XOXO (or any specific format) for > now. These two options are the same for any representation format we > choose to go with: Either we have a MIME type or we don't. If we don't > have a single consistent MIME type, MIME dispatching for introspection > documents is thrown out the door. >
Why "single consistent"? What is the benefit there? I think there is great value in being able to serve documents that can be browsed with one Content-type header and kick off some other app with another Content-type header. Why maintain two documents with the same info when one would do? - Luke
