[IMXPwg] next rev of drafts, incomplete

Graham Klyne GK@Dial.pipex.com
Wed, 20 Sep 2000 08:42:49 +0100


At 10:16 PM 9/19/00 -0700, Marshall Rose wrote:
>
>2. i question whether using rfc2533 really yields capabilities info or 
>not. darren doesn't think so. or rather, he doesn't think that it yields 
>the stuff that you want in terms of capabilities (e.g., "i accept text/html").
>
>graham - can you clarify?

Hmmm... not quite sure what needs clarifying here.  The short answer is 
"yes, it should".

RFC 2533 is a format for describing receiver capabilities and 
sender/document characteristics.  There is no carrier protocol defined 
there.  The "I accept" bit has to be implied separately.  The same basic 
format can be used in different ways.

So, if we publish an RFC2533 expression as capabilities in a presence 
payload, that describes the receiver capabilities.  For example, receiver 
capabilities published as:
     (type="text/html")
means the receiver can accept MIME content-type text/html.

The expression:
     (| (type="text/plain") (type="text/html") )
means the receiver can accept text/plain OR text/html.  This case can also 
be written as:
     (type=["text/plain","text/html"])

A slightly more complex example:
     (| (& (type="text/plain") (color=binary) )
        (& (type="text/html")  (color=limited) ) )
means the receiver can accept text/plain without colour, or HTML with 
limited colours.

We've used this stuff to describe the capabilities of various kinds of 
group 3 fax machines, which can really get quite complex.  The expressions 
reflect that complexity (unsurprisingly).  See RFCs 2879 and 2880.  Also 
note RFC 2913 ('type' tag) and RFC 2534 ('color' tag, and some others.)

Finally, I've done an I-D (expired now, I think) with an improved and 
simplified description of a feature matching algorithm.  There's also some 
Java code at the 'medfree' page on the IMC web site.

>IV. imxp-message
>
>this is just skeletal.
>
>graham - do you want to work on this?

I'll take a shot.  It will be competing for time with other activities, so 
I won't achieve your rapid-fire turnaround.  Can you send me a copy of your 
XML source for one of the existing drafts (with all the authors, 
references, etc. in your XML format) that I can use as a start point?

#g

------------
Graham Klyne
(GK@ACM.ORG)