We need to start talking about group and public annotations

Brian Behlendorf (brian@wired.com)
Wed, 20 Jul 1994 02:39:00 +0200

(RPG - I'm cc'ing you coz I saw your post on c.i.www about this and
want to bring this topic to this forum)

Okay, gang, I've looked EVERYWHERE I could, I've been on www-talk for the
past 8 months and I've gone to the web conference in Geneva, and yet this
topic eludes my research into its current status.

The problem: we need to bring back group and public annotation BACK to
the WWW.

Annotation is a meme that has been around since day 1, as far as I can
tell. It was in NCSA's early server and browser, but they removed it
as I suppose there were other things they wanted to get working fine
first. (Don't blame that, I guess). Support for user annotations is
still in Mosaic, but I don't think it was ever implemented in any other

I found *some* documentation on NCSA's server about annotation, but no
implementation details or underlying structure. I figured it was tied into
the PUT method, but I haven't been able to find much on that either (I looked
around CERN's server too). It's so frustrating because the interface is
*almost* there in NCSA's Xmosaic, it's just the "group" and "public" option
is greyed out! :)

For Wired's new online publishing effort launching in September, I am writing
a collection of CGI scripts and forms, essentially an extension of WIT, that
will allow users to give feedback on any document on our site, and for vines
of discussions to cascade from that feedback, also implementing a peer-review
system which will make navigating the feedback and finding useful commentary
easier. It'll be something like WIT, but structured differently. I believe
commentary functionality should be built into the client and server, the way
it is in X Mosaic, and that the PUT method should be enhanced and solidified,
instead of trying to kludge this as CGI scripts and forms. It should also
thus be standardized across servers so we don't have to relearn
administration details for this for a different client, and so that
subprograms can work together. Part of implementing this involves a huge
database of users with posting authority, and writing tools to manage that
database and the verification issue.

So far I've been planning on doing this through the CGI/forms interface,
but I want to do it the RIGHT way, which would be annotation as
implemented in the Mosaic browsers and using PUT. I also would want to
work with the WWW community on this project, because there are others who
seem to be interested in this as well. I also don't have a problem with
putting the work I do in the public domain.

I am working on a spec for the system and a layout of the different
scripts and directory/file systems. What I want to know now is, is
this a topic we should keep discussing here, or is the body of people
interested in working on this sufficiently small enough that we could
break up amongst ourselves and create a sublist.

This is also a call for information. If someone could point me towards
documentation for the PUT method and for any more info about implementing
annotation, let me know.