Re: Where's

From: Lee Daniel Crocker (
Date: Tue Feb 15 2000 - 11:03:50 MST

> Since the topic has gotten around to the crit service, and it seemed
> IMHO to be one of the better designed of such systems (I now know of
> about 5, most used for graffitti), I've been wondering if anyone had
> any comments on its design. What does it do right and what does it do
> wrong? How would you change it, and what would be your wish list if
> you didn't have to worry about implementation details?

None of these services can possibly succeed or make any difference as
an independent service, least of all Crit because they have no realistic
business model. There is a way to get annotations working and accepted,
though, and that's to make them part of the infrastructure. The Mozilla
project (which everyone mistakenly believes to be a failure when it has
made a huge impact and will make a far larger one) provides an opportunity
for that (JWZ has already suggested this, but I don't know if anyone's
actively working on it).

Very few of you will remember that the original NCSA Mosaic web browser
(from which Netscape evolved) had a user-annotation feature that stored
user comments on a central server. This feature was removed because it
was felt (correctly) that this server would become a bottleneck and
become clogged with lots of hard-to-maintain data. What should be done
is to create a standard annotation service interface protocol that all
the competing services understand, and add user-level support for the
protocol as a standard part of the browser.

Here's the user's view: from a web page (seen purely as intended by its
author with no Crit-pollution), the user selects the "View Annotations"
menu or toolbar button. A list shows "31 comments at, 12
comments at, 5 comments at", etc. The
user selects one of those links, and brings up one of
its pages--drawn exactly as it wants them drawn, possibly with ads or
other gadgets, that contains those comments. User's comments can be
added from those pages, using whatever means the annotation service
provides (they'll want to be different so they can control formatting
and content and provide value--and yes, the comment servers can and
should exercize their own editorial control, which might be none for
some services). The browser will ship with a list of well-known
annotation services that support the protocol, and the user will be
able to add more or remove ones he doesn't want.

I could hack a protocol in a day or two--beyond that we'd need
commitment from one or more services to support it and a few folks
to put the code into Mozilla. Netscape 6 will be based on Mozilla,
as will other browsers. If the feature catches on, MS will have to
add it (and there's little reason for them not to--in fact it's an
opportunity for them to add their own censored annotation service
as well). After that, it's just part of the Web.

The critical keys to success are that PROTOCOL should be open and
free, while CONTENT must allow editorial control, or else it will
just be a worthless commons.

Lee Daniel Crocker <> <>
"All inventions or works of authorship original to me, herein and past,
are placed irrevocably in the public domain, and may be used or modified
for any purpose, without permission, attribution, or notification."--LDC

This archive was generated by hypermail 2b29 : Thu Jul 27 2000 - 14:03:47 MDT