Event delivery as part of web architecture?

I'm prompted to write this (in part) by an announcement [1] of Distributed
Multimodal Synchronization Protocol (DMSP) activity to the IETF application area
general discussion list [2].  That message refers to an Internet draft about
DMSP [3].  I also note that DSMP relates to W3C work on multimodal systems, so I
thought it maybe worth mentioning here some of thoughts I mentioned on the IETF
list [4].

My recent experience leads me to believe there is a gap in the Web protocol (or
"interaction", to use the language of AWWW [6]) landscape.  I've recently come
across three very different applications, all of which are developing different
solutions for what I see is fundamentally the same underlying requirement, viz.
delivering notification events to a web browser:

 1. DMSP, already mentioned.  This requirement is plainly illustrated on page 7
of the Internet draft [3].

 2. Portal software, where there is a desire to synchronize the views of
different portal channels.  Currently, the portal community is developing
specifications for inter-portlet communication [7], which I regard as an
overweight solution to the problem they are trying to solve.  I've written some
thoughts about this at [5], in particular describing how I think a simple
mechanism for delivering notification events, and how such a solution would be
useful to more than just portal software.

3. My own work on "smart home" control systems based extensively on Web
technologies.  I am using Javascript/Ajax techniques to implement a simple event
delivery notification so that browser displays can automatically update to
reflect changes in the controlled environment.

Beyond these three specific areas where I think event notification would be
useful in other areas, such as RSS/Atom feed handling, displaying dynamic web
resources like stock prices, etc.  In my wiki pieces at [5], I outline what
seems to me to be a plausible strategy for defining and introducing such a
capability, using Javascript as a stepping stone, particularly directed to the
context of multi-view synchronization for portals, but also more widely
applicable.  At the heart of this proposal is the principle of specifying a very
simple event delivery protocol mechanism (probably an existing protocol) from
the event descriptions themselves (e.g. as in email, with its separate
specifications for SMTP and RFC2822/MIME).

To the extent that this affects (extends) the discussion of interaction in
"Architecture of the World Wide Web, Volume One", is this something that is
appropriate for TAG to consider and/or try to exert some influence over?

#g
--

[1] http://www1.ietf.org/mail-archive/web/discuss/current/msg00275.html
[2] mailto:discuss@apps.ietf.org
    http://www1.ietf.org/mail-archive/web/discuss/current/index.html  (archive)
[3] http://www.ietf.org/internet-drafts/draft-engelsma-dmsp-01.txt
[4] http://www1.ietf.org/mail-archive/web/discuss/current/msg00276.html
[5] http://wiki.oss-watch.ac.uk/InterPortletCommunicationConsideredHarmful
[6] http://www.w3.org/TR/webarch/
[7] http://www.oasis-open.org/committees/tc_home.php?wg_abbrev=wsrp
    Cf. "Cross-Portlet Coordination SC: Chartered to define how portlets can
react in a coordinated manner to user interactions with any portion of the
overall application."


-- 
Graham Klyne
For email:
http://www.ninebynine.org/#Contact

Received on Sunday, 12 March 2006 18:13:17 UTC