[ChicagoLinux] Programming Meeting Retrospective

[John Quigley]

Jordan Wilberding wrote:
> So this project is in all python?

Yea, we made that executive decision on Friday.  Hope that's cool.

> Also, could we get a better definition of what this is? I am not
> familiar with the term feed-mesh web application. While I could guess, I
> would rather have an answer from someone that worked on it. :)

Let's see if I can properly commit my thoughts to 'paper'.  We want to 
extract what's good and interesting from the web.  Feeds are a good step 
in that direction: they provide a thin layer of abstraction over a web 
page (or set of web pages, as the case may be), and help in condensing 
information.  Let's build on that.  The motivation is to fix what reddit 
got wrong.

Imagine a simple personal feed aggregator, something not unlike Google 
Reader.  We provide a web site.  It allows a user to log in, identify 
feeds they're interested in, and these are displayed on their personal 
page.  Nothing new here.

The next step is to begin tracking trends in these feeds, specifically 
what links people visit.  These trends will be reflected in what is 
presented on the application's main page.  Links will migrate up on the 
front page according to the number of people who click on them in some 
time frame, and start to sink using some decay algorithm.  We can get 
extremely fancy with this, going so far as to index the pages that 
people visit in an attempt to gather semantic details about what an 
individual likes.  Lots of potential to stack functionality in a modular 
way.

We want to provide the ability to create ad-hoc groups.  Let's say you 
and I are friends, and we want to create a 'programming' feed mesh.  We 
tag each of our respective feeds that are 'programming' oriented, and we 
can visit a group page that displays links according to trends in our 
visitation of links in this feed grouping.  This is like flexible, 
user-driven 'sub-reddit'.  This, likewise, can get very fancy.

Now, there's the potential to track and trend vast amounts of data with 
this, which would be interesting from our perspective, but the users may 
not be too cool with that.  What to do?

Open source the platform/application and allow anyone to setup their own 
service.  Many feed mesh nodes can pop up, all over the web, and users 
can control their data.  We can network these independent meshes 
together, if the site admin wants.  I'm building out a custom HTTP 
server so that I can extend the protocol to support sharing of data 
collected.  We could provide the 'mothership' application that would be 
the default location where this information is collected.  Hopefully 
that will allow us to present very relevant, interesting stuff extracted 
from the web.  This is like a planet on steroids.

> Sounds like you guys made a great start though! The biggest thing now is
> to make sure it keeps on going. I will do my best to help out!

The intention is to constrain our scope to the most simple application, 
get that out the door, and begin adding components (functionality) in a 
modular way over the summer.  I think the five of us who met this past 
weekend, if no one else, are motivated to make this happen, if only 
because we're dying to write more code.  I think this will be a success. 
  Looking forward to your help =)

- John Quigley