RE: Section 3.5. Passing fragment identifiers to other systems.

I'm not sure I'd call it "abuse". But when moving a document
from Draft Standard to Standard, it's reasonable to restrict
backward incompatible changes; if this is the "legacy"
view, so be it.

I believe that it has been legitimate for URI implementations
to assume that the main URI can be separated from its fragment,
handed off to a separate URI access mechanism (which looks at
the scheme and does scheme specific processing) and then
the fragment is applied after the results have been accessed,
without reference to the scheme or any of the other components
of the URI.

This isn't "http" specific, because it works across access
schemes http, ftp, file, data, cid, 

Other kinds of processing, hinting, and uses of fragments
during processing might be _allowed_ and even reasonable, but
*requiring* that the scheme-specific implementation can access
the fragment (by defining the scheme-specific fragment interpretation)
should not be allowed.

Larry

Received on Friday, 27 February 2004 20:18:05 UTC