Subject: [NYCGA Internet] new project - ows news curator
From: Sam Boyer
Date: Thu, 20 Oct 2011 08:10:31 -0700
To: ows_solutions <>, Internet Working Group <>
CC: Dave Winer <>, Thorin Caristo <>

hi folks,

((tl;dr: we've built a prototype for an aggregated content curator. the
initial focus is on rolling it out for liberty plaza; rolling it out on
demand to solidarity occupation sites is not too difficult and on the
horizon. we need help with final IA plannign & design implementation.))

i'm a bit ashamed to admit that i first started coming down to liberty
plaza only at the beginning of last week...but hey, better late than
never. the only remotely tech-ish looking people i could find in the
park were the folks clustered around the media station, so i made
friends with them, particularly Thorin. i tried to suss out how i could
best help with my background as a software engineer/drupal guy, and our
conversations eventually led to the idea that they could really use a
central aggregation platform through which they could embed the
livestream as well as pull in news, videos from across the net via feeds
(primarily, but not exclusively, RSS), curate by choosing a subset of
the articles to actually display, then roll it all into a nice-looking,
public-facing interface.

this basic workflow is a *very* close match to the functionality
provided by the managing news ( drupal distro.
so, while i was in portland over the weekend for the pacific northwest
drupal summit, i rallied a bunch of support around speccing out and
building a MN-based system that would meet with these requirements. and
we built out most of it, though we ran into some problems. we ended up
deciding to built it in two separate components, a frontend site (fairly
vanilla drupal7 + views) and a backend site, powered by managing news
connected by pulling the rss feeds into the frontend from the backend,
because it was easier to build a custom user-facing presentation layer
than hack the tightly-coupled theme that MN ships with. or at least,
that's our estimation, we haven't actually really *done* that yet - i
have a surplus of engineers but a shortage of frontend folks at the
moment :(.

in any case, the backend interface is visible at . and the workflow goes
roughly like this:

1) Add a feed from which to pull in content for curation. (We started by
using the OPML that Dave Winer has been keeping as a central aggregation
point -
2) Create a channel into which items can be curated. Channels can be
configured to watch a tag (or tags) on articles that have been pulled in
via feeds.
3) Get curatin'. Since channels watch all articles for sets of tags, the
entire curation process is as simple as tagging/untagging a particular
article. Editing an articles tags can be done directly via the feeds
interface, seen here:
(if you're logged in). Note that this is really a portal site, a bit
like google news, in that it does not contain any original content; when
you click a link there, it
4) In the normal case with MN, you use the built-in interface to
drag-and-drop a bunch of blocks, each of which aggregate something (a
channel, a feed) in some visualization form (a plain list, a geo
representation, etc.).

the complication is that the the IA of MN is really not what we need. we
want to *only* show a very targeted interface to the general public: a
frontpage which combines a number of channels together in some fashion,
and then individual pages for each of the channels. we don't want that
nav at the top that MN provides, we don't necessarily want end-users to
get lost in the sea of thousands of articles from all the raw feeds
themselves, etc. And, given that this is a devseed joint, the system is
great...until you want to customize it. hence the separate frontend
site. so step 4 is a bit different as frontend management takes place on
the other site, but the truth is that most content curators will spend
90% of their time in steps 2 and 3.

the real win with this system is that we can put these sites in the
hands of highly non-technical people in physical occupation spaces,
allowing each site to massage the massive pipe of #ows-related news into
a picture that they feel best pertains to, captures what's happening in
their particular space. and with all the wonderful agility of being
feeds-driven, so content can basically be grabbed from anywhere. seems
to me this makes it a bit of a 'holy trinity' - a) entirely
feeds-driven, so builds perfectly overtop of our existing focus on
aggregation b) curation entirely by non-technical folks, which could be
very empowering for occupiers c) pretty easy to replicate for any
occupation space that wants one.

this was all pretty much done by sunday afternoon. the place that we
sorta got stuck, though, was on designing the frontend. we had a bunch
of engineers show up at the bof to work on it, but were light on
designers; beyond that, i didn't have a very clear end-user interaction
experience in mind (i never did clearly define the audience with Thorin)
due largely to some fuzziness around the exact way we expect this system
would be used. we did create a basic IA + design, and it's supposed to
be up at , but it seems
we're having some trouble getting it to show up properly. should have
that resolved today.

there are definitely some interesting additional features we've
discussed - for example, creating bookmarklets for curators that utilize
opengraph in order to folks just click the bookmarklet in order to make
the article available in a special feed, so that stuff can be added
without having to find an RSS feed. but those are outside the
minimum-viable-product scope, so we're putting them off for now.

i've created
to track work on this tool. i'm looking for a project manager for this
(though i may have found one just today), and we definitely need
designers (though i had five respond to my tweet last night). however,
if you think the general direction this is going is interesting, PLEASE
get in touch with me so we can discuss the right space/application for
this tool. at this point, i think what we really need is more "product
owners" (to use the agile terminology) than anything else. folks willing
to help transcribe some of these ideas into the redmine project wiki
would also be SUPER-welcome - hell, it took me three days to even get
this thing written up.

i suspect there are wide-ranging applications for this - for example, a
packaged version of this system could be a significant part of, or even
an entire answer to, some of the needs being expressed by the OWSJ here .

also to be clear, i did very little actual work on this except vigorous
handwaving and a bit of structure planning - that was really primarily
dave tark and andy laken, though there were at least half a dozen -
sorry folks, i don't remember everyone who was there now offhand :(. and
clearly we'll be adding some more folks to the team as we go :)