future of mosaic annotations
marca@ncsa.uiuc.edu (Marc Andreessen)
Date: Tue, 31 Aug 93 02:05:13 -0500
From: marca@ncsa.uiuc.edu (Marc Andreessen)
Message-id: <9308310705.AA09208@wintermute.ncsa.uiuc.edu>
To: www-talk@nxoc01.cern.ch
Subject: future of mosaic annotations
X-Md4-Signature: 9a091f95f1a6a44014b209b48f4cce3e
Status: RO
We're thinking of ripping the *existing* group annotation support out
of the Mosaic 2.0 client and upcoming revisions of the NCSA httpd, and
shutting down the existing annotation server on hoohoo.ncsa.uiuc.edu.
The reasons are:
(a) The existing protocol was strictly experimental, and at the recent
developer's conference, a better, HTTP/1.0-based annotation
protocol was devised.
(b) The existing group annotation server software is sadly deficient
in many ways, including notification and annotation database
management; a rewrite will be required to fix these problems (they
have proven nearly crippling for real use of group annotations).
(c) The annotation server on hoohoo (which warped the original idea of
"group" annotations by being public to all Mosaic users) has
outlived its usefulness as a global mechanism and should be put to
sleep.
The first instantiation of the new annotation protocol will likely be
in the form of "public" -- that is, annotations that reside on the
same server as the documents they annotate -- annotation support, as
opposed to group annotation support. (This is one result of the
hoohoo experiment...) A future version of Plexus will likely support
this (not to overcommit Tony :-). Public annotations should be *much*
more useful to a larger number of people than the existing scheme --
the solution is scalable and everything should be faster.
The same (new) protocol should also be relevant for the group
annotation principle, and a future version of the NCSA httpd will
probably include group annotation support based on this protocol (with
the underlying current problems of notification and annotation
database management addressed).
How's that sound?
Marc