[annotator-dev] Looking at Fluidinfo as a possible backend for Annotate

Kartik Subbarao subbarao at computer.org
Wed Aug 1 15:41:11 UTC 2012

On 07/26/2012 05:08 PM, Kartik Subbarao wrote:
> Cool! I have a post submitted to the fluidinfo-discuss mailing list
> proposing a basic design and asking for some refined recommendations,
> I'll see what I hear back.

An update on this -- I've gotten some good information/recommendations 
from the fluidinfo team, and am working on a first-pass implementation.

Also, it turns out that fluidinfo is working on a comment aggregation 
site, which could be a good opportunity for collaboration. Here's the 
relevant excerpt from the email, I'd be curious to hear your thoughts:

So far we've been talking about storing comments as tag values (along
with the other metadata that goes with the comment text), which is
generally a good approach.  We've recently been working on an
application called loveme.do which aggregates comments from social
networks about hashtags and URLs and presents them in a nice uniform
way.  For example,


Probably most of the comments you see there will be from Twitter, but
they can also come from Disqus, Tumblr and Facebook (and in the
future, other services).  It could be interesting for Annotator to be
one of the services that's pushing comments into Fluidinfo.  In that
case, we create an object for each comment and we use a custom (so far
internal) API for creating them (which does some special analysis and
linking of the content).  If this is interesting we could talk more
about how that API works and you could add your additional metadata to
the comment objects created by that API.  A way to think about this is
that objects have two kinds of data: structured tag/value and a
comment stream.

More information about the annotator-dev mailing list