Re: Several Content-Encodings

I think the problem here is language. As a general rule the order in
which differently tagged headers appear is not significant, the order in
which tags of the same name appears is significant. A proxy may alter the
order of differrently named tags but not the order of tags with the same
name.

This area impinges on the PEP proposal, PEP is in part intended to solve
this problem. Consider a document which is compressed, signed and encrypted
in that order. It is essential that the transport prserves the information
that this is the order in which the transformations occurred.

The content encoding model is unfortunately limited. I don't think we should
try to fix it now. I have a feeling that Jim and Larry would be upset if we
were to start.


		Phill

Received on Wednesday, 10 April 1996 10:16:18 UTC