The Web community has a broad agenda for the future of HTTP. Users, developers, and webmasters have projected their hopes and dreams onto several hypothetical directions for HTTP. Over the next year or so, we expect significant, concrete progress on assessing the limitations of HTTP/1.x and designing extensions to address them. Some of these areas include performance engineering, distributed authoring and versioning, asynchronous notification, distributed object interfaces, tuning for embedded Web devices, and realtime multimedia support. Participants are also encouraged to submit position statements on other aspects of the future evolution of HTTP: requirements, deployment issues, standards process issues, and lessons learned from developing similar protocols.
WWW7 offers a unique, neutral forum to discuss these issues within the Web community, separately from its development and standardization processes.
Submissions may be considered for inclusion in "Transfer Protocols: Plumbing the Web," an upcoming issue of the World Wide Web Journal, published by O'Reilly & Associates.
Please see the Abstract.
The meme of a new generation of HTTP has been around for a while. New application semantics, encodings, protocols, bandwidth management, and object models have already been projected onto this canvas. This workshop will explore these influences along with many other technical, economic, and political issues.
The IETF WebDAV working group is an existing effort working to define the future of the Web. WebDAV, World Wide Web Distributed Authoring and Versioning, is working to extend the HTTP/1.x protocol to support remote, asynchronous, collaborative authoring of Web documents. WebDAV provides an infrastructure for creating locks, recording properties on resources, collection and namespace operations, and simple nonlinear versioning. Lessons learned from this effort will inform the development of HTTP-NG.
Simplifying HTTP implementations and multiplexing concurrent connections were the twin goals of Simon Spero's 1994-5 era Session Control Protocol (SCP) and ASN.1-encoded binary HTTP-NG. Both aspects of his proposal have been quietly explored in the research community while HTTP/1.1 standardization was in the foreground.
The broader TCP community has also been coming to grips with the outsize effects of HTTP design on Internet performance. Several proposals have addressed these issues before, notably Transaction/TCP and current Internet-Drafts on recommended implementation behavior. The IETF Transport Directorate has placed a high priority on coping with the challenge of Web traffic.
Yet another which has been woven into dreams for an "NG" is a distributed object system. The Object Management Group has been enthusiastic about merging HTTP and IIOP. The idea is that GET, PUT, POST of the HTTP/1.x object model can be viewed as an RPC system for documents -- what if NG evolved into a dynamic invocation interface for any class of object? Even the fundamental mission of HTTP as a HyperText Transfer Protocol is up for renegotiation.
In addition to research work and related IETF efforts, the W3C officially launched an HTTP-NG Activity on July 21, 1997. That group will be investigating issues (focusing on simplicity and extensibility) relevant to the next generation of HTTP, led by Henrik Frystyk Nielsen and Jim Gettys. In August 1997, IETF had a BOF session for HTTP directions in Munich as well. The World Wide Web Journal is also planning a special issue focusing on the next generation of Web Protocols for early 1998. We believe the HTTP-NG development process is beginning to coalesce and the WWW7 conference in April 1998 is an ideal point to discuss technical and political influences on the future of HTTP.
References and pointers to relevant papers:
We anticipate this workshop will appeal to a wide range of participants, such as:
We expect to report on the results of the workshop on Developer's Day. Selected position papers will be considered for inclusion in a special issue of the World Wide Web Journal on "Transfer Protocols: Plumbing the Web."
This workshop requires an overhead projector for slides, and an overhead video projector for computer display. Internet connectivity would be strongly preferred.
We intend for this to be a full-day workshop co-chaired by: Rohit Khare <rohit@uci.edu>, Henrik Frystyk Nielsen<frystyk@w3.org>, and E. James Whitehead, Jr. <ejw@ics.uci.edu>.