Re: a comment about paper

Thanks for the comments- I'd only heard the quietest rumblings about PICS-
it sounds salient, and I'll go read more.

>
> Also, the scaling of large annotation systems into the already
>crowded internet is problematic.

I agree that it's problematic. As you've probably gathered, I also think
it's critical. 

> Some solutions have been
>suggested for improving the performance of HTTP (http 1.1 being
>one). There can be no ONE POINT for housing all the public comments
>about a given resource.

I suspect that I'm way less technically savvy than most people on the list,
and am thus a little wary about arguing feasibility. I recognize that a
central public comments database would be huge- definitely too big for a
single publicly-minded individual or organization to maintain out of
altruism or in the name of research. But it's not clear to me that it would
be bigger than existing free commercial web search engines- and it does seem
possible to operate it on a similar business model. (Assuming, of course,
that Web search engines actually are profitable- for all I know, they're all
still losing money...)

Are there points I'm missing there? There's some stuff in the paper where I
take some really rough stabs at imagining how much processor
power/bandwidth/storage the project would take relative to running a big
search engine. I'd love to know if there are complications I'm missing...

I also wonder about running it on a distributed model a la Usenet, which is
a pretty big whack of data to which anyone can  write or read (with no one
point for storing/serving all that info). I'm less optimistic about this
possibility, though...

>
> In one way, standalone comments about a particular site are a simpler
>version of annotations. in general, an annotation will carry with it
>a collection of hooks that have to be dropped into a particular
>version of a target resource. This is a much more complex issue
>than simply grafting annotation to the end of a given resource.

Agreed. From what I've read about the w3c-annot ideas, it's a way richer
medium than the simple comments that I talk about in the paper- lots of very
cool tools that everyone would love to see implemented.

I wanted to describe a very simple commenting structure, in order to focus
on the specific idea of trying to make the commenting (or annotating) system
as universal as possible- to make it possible for me to see every comment
that's been added to a page, and to add comments that'll be similarly
available to others. I don't know if there are extra complications that
would arise in a richer system closer to the w3c-annot one.

>
> Another missing piece of the puzzle is... what if the underlying (host)
>resource changes after a set of (parasitic) annotations have been applied 
>to it... and what if the changes are so drastic that the context of the
>original annotations are lost. Obviously it is the right of anyone
>to change any URL response at any time (or is it???), but wouldn't
>you like to be able to keep your snarky comments in line with the 
>host resource? This is screaming for some sort of notification
>system which has knowledge of the web of annotations and the 
>responsible authorities for each one. 

Yeah.I'd thought about this a little.  Say  "Steve's page of spurious
trash", at http://www.interweb.com/~steve/trash.html, collected a huge
volume of critical comments. Steve moves it to a different ISP and
distributes it accross different pages. The comments it garnered needn't be
lost. Only one user would have to post messages on the new pages desribing
the move and pointing to
http://www.annotserver.org?http://www.interweb.com/~steve/trash.html. (Or
whatever the syntax would be). All the old comments would then be availble.
Aince comments are maintained independently of the host, there's no reason
an url has to exist to support comments.

>
>That's probably only worth about half a cent.

It's half a cent more than I had before. Thanks.

- Misha Glouberman

Received on Tuesday, 30 July 1996 11:18:43 UTC