- From: Mark Nottingham <mnot@akamai.com>
- Date: Mon, 16 Apr 2001 23:10:19 -0700
- To: XML Distributed Applications List <xml-dist-app@w3.org>
It looks like the consensus there is "or not"; I was too quick on the forward. There was fairly decisive discussion of the issues (I'd like to point to an archive, but I can't find one that's current; if you're really interested, mail me). Basically, the fact that MIME parsers would have to be substantially rewritten, coupled with the feeling that this can be better addressed at a different layer (e.g., transport, with BEEP) or other techniques, means that this isn't considered a practical approach. On Mon, Apr 16, 2001 at 03:14:54PM -0700, Mark Nottingham wrote: > > > FYI; may be interesting in the context of SOAP with attachments for > small devices (or not). > > > ----- Forwarded message from "Herriot, Robert" <Robert.Herriot@pahv.xerox.com> ----- > > Date: Mon, 16 Apr 2001 12:36:12 -0700 > From: "Herriot, Robert" <Robert.Herriot@pahv.xerox.com> > To: discuss@apps.ietf.org > Subject: Two new drafts: Multipart/Interleaved and Application/BatchBeep > X-Mailer: Internet Mail Service (5.5.2653.19) > > I recently published two drafts > > > http://www.ietf.org/internet-drafts/draft-herriot-multipart-interleaved-00.t > xt > > http://www.ietf.org/internet-drafts/draft-herriot-application-batchbeep-00.t > xt > > (Note that the first draft is not available via the Internet Drafts Search > engine, but is accessible via the URL supplied above). > The problem that motivates these two proposals is the case of an XML > document with attachments, such as gif images. Normally, a Multipart/Related > representation would suffice. The first body part would be the XML and the > remaining body parts would be gif images, each referenced from the XML. But > when a memory constrained printer encounters a reference to a gif image body > part in the XML, it might not have the space to hold the remaining XML while > it scans ahead for the gif image. This problem suggests the need to be able > to interleave chunks of XML and gif images. Each of the two drafts provide a > solution for interleaving chunks. > Both drafts define new MIME types. The first draft defines > Multipart/Interleaved and Application/Chunk content types, which extend > Multipart/Related for use with memory constrained devices. The second draft > defines Application/BatchBeep content type, which defines a MIME type whose > contents is the client-to-server portion of a BEEP session. Although the > two drafts are based on two very different ideas, the examples show that the > two solutions are nearly byte-for-byte identical. The primary difference is > that the boundary string in the Multipart/Interleaved solution is a trailer > and/or header in the Application/BatchBeep solution. > I would appreciate any comments on these drafts. > Bob Herriot > -------------------------------------------------------------------- > The following is the abstract for the Multipart/Interleaved draft. > The Multipart/Interleaved content-type, like the Multipart/Related > content-type, provides a mechanism for representing objects that consist of > multiple components. Each body part of a Multipart/Related entity > represents a component, whereas each body part of a Multipart/Interleaved > entity represents either a component or a part of a component. A body part > that represents a part of a component has the content-type of > Application/Chunk. With Multipart/Related, a body part and its reference in > some other body part may be separated by many octets -- more octets than a > memory-constrained device can deal with. With Multipart/Interleaved, a body > part can represent a part of a component. This allows a body part and its > reference in some other body part to be made quite close -- close enough for > a memory-constrained device to deal with. This document defines the > Multipart/Interleaved content-type and the Application/Chunk content-type. > It also provides examples of its use. > --------------------------------------------------------------------- > The following is the abstract for the Application/BatchBeep draft. > The Application/BatchBeep content-type, like the Multipart/Related > content-type, provides a mechanism for representing objects that consist of > multiple components. An Application/BatchBeep entity contains the wire > representation of the client-to-server part of a BEEP (Blocks Extensible > Exchange Protocol) session, which consists of sequence of frames. Each frame > contains a message or a part of a message. Each message (other than channel > 0 messages) represents a component of the compound object, just as a body > part of a Multipart/Related entity represents a component. With a > Multipart/Related entity, a body part and its reference in some other body > part may be separated by many octets -- more octets than a > memory-constrained device can deal with. With an Application/BatchBeep > entity, a frame can contain part of a message. This allows a message and its > reference in some other message to be made quite close -- close enough for a > memory-constrained device to deal with. This document defines the > Application/BatchBeep content-type. It also provides examples of its use. > > ----- End forwarded message ----- > > -- > Mark Nottingham, Research Scientist > Akamai Technologies (San Mateo, CA USA) -- Mark Nottingham, Research Scientist Akamai Technologies (San Mateo, CA USA)
Received on Tuesday, 17 April 2001 02:10:31 UTC