Re: Transport level compression

Arvola Chan wrote:
> 
> Thanks Chris.
> 
> My comments are inline.
> 
> -Arvola
> 
> -----Original Message-----
> From: christopher ferris <>
> To: Arvola Chan <>
> Cc:  <>
> Date: Tuesday, August 21, 2001 1:11 PM
> Subject: Re: Transport level compression
> 
> >Arvola Chan wrote:
> >>
> >> HTTP allows transport level compression through the use of the
> >> Content-Encoding entity-header. The permissible compression
> >> algorithms include gzip, compress, and deflate.
> >>
> >> Should the use of transport level compression be specifiable in
> >> the CPA when the transport is HTTP?
> >>
> >> Thanks,
> >> -Arvola
> >>
> >> ----------------------------------------------------------------
> >> To subscribe or unsubscribe from this elist use the subscription
> >> manager: <http://lists.oasis-open.org/ob/adm.pl>
> >
> >Arvola,
> >
> >This is an interesting question. I would say that this is probably
> >a good idea. However, we shouldn't limit to Content-Encoding.
> >Seems to me that what a CPP/A should define (at least for HTTP)
> >is equivalent to the HTTP Accept* request headers that a client
> >uses for HTTP requests to inform the origin server what its
> >capabilities and/or preferences are.
> 
> <arvola>
> I agree in principle.
> </arvola>
> 
> >
> >It also seems to me that one could apply the Accept headers
> >abstractly to any transport protocol. Thus, something like:
> >
> ><c:Transport transportId=".." xmlns:c="...">
> > <c:SendingProtocol version="1.1">HTTP</c:SendingProtocol>
> > <c:SendingProtocol version="1.1">SMTP</c:SendingProtocol>
> > <c:ReceivingProtocol version="1.1">HTTP</c:ReceivingProtocol>
> > <c:Accept c:mediaType="..."/> <!-- zero or more -->
> > <c:AcceptEncoding c:contentCoding="..."/> <!-- zero or more -->
> > <c:AcceptLanguage xml:lang="..."/> <!-- zero or more -->
> > <c:AcceptCharset charset="..."/> <!-- zero or more -->
> > ...
> ></c:Transport>
> >
> 
> <arvola>
> Unfortuantely, contentCoding is applicable to HTTP only. SMTP /
> S/MIME has content transfer encoding but that is limited to 7bit,
> 8bit, binary, quoted-printable, and base64.
> </arvola>

I am not being specific to the headers specified in HTTP, just the
concepts that they embody which I believe are somewhat universal.

> 
> >Comments?
> >
> >Chris