TTI Vanguard Conference Notes - 3

Lili Cheng, Future Social Experiences (FUSE) Lab, Microsoft
Designing Applications and Services for Our Social Nature

We asked people in a mall who they interact with, they would put themselves at the centre and different groups of people in clusters around them. So we tried to do was to do the same sort of thing with your email, and it builds a similar sort of cluster map for you. This theme of self-organizing and origanizing things in a gorup for you is something you'll see in a lot of our work.

This team built the matchmaking behind xBox 360, Bing search stuff, Bing Twitter maps, Windows desktop search index, aggregation, etc.

Example: Bing social search, called 'Twig'.  See http://www.bing.com/social - it is a good breaking news service.  (Some other examples - Montage, Spindex - URLs are obscured in IE display so it's hard to show them - also you have to have a Windows Live ID to use Spindex). Annoying.

Example: Inner Circle. Based on the idea of 'wallop' applied to email. And we wanted to bring social information into email - view of 'Salsa', which shipped with Outlook.

(Discussion of Microsoft vis-a-vis Facebook, whether people use Sharepoint, its relation to open source, etc.)

Docs.com for Facebook - we wanted to explore 'what happens when work software moves out into other areas'. We worked with Facebook, they gave us access to their APIs. It was interesting to see how people in Facebook use docs. It was modeled after the photo app. One thing we discovered is that people don't really want to create docs in that way - so we created a document that will create themselves - eg. an excel doc that is created from your friends graph.


Brian Behlendorf, Apache Software Foundation
Peer Production and Public Policy

Talking about trying to encourage government and the private sector to work together to solve some of the big problems of the world.

Obama's open government initiative memo - order to make government more transparent, participatory and collaborative. Led by Beth Noveck, the idea was to create a document the president could sign actually created by the participants. The website is http://opengov.ideascale.com/

People decided that policies related to illegal immigrants were in scope, that the birther issue was a transparency issue, etc. And we realized that there was maybe something amiss. A bunch of issues, very little to have to do with transparency, were promoted to the top.

But the next tier had some real ideas. Eg. whistleblower protection, open CRS reports, online interactive federal budget website, stand-alone (no earmarks) bills. Etc. These then went back into the bureaucracy to be redrafted into forms that could actually be implemented.

Since then, HHS has been active with http://www.hhs.gov/open/ See also Sunlight labs, 'Design for America', which uses the HHS data set. One of the winners of the contest was the 'county sin' rankings.

What did we learn from this? Here's a set prepapred for this presentation:
- we weren't driving toward specific outcomes - just created a way for flash mobs to vent
- transparency from day one in processes and assets
- recognition of participant motivations - a lot of people participate in government because it's part of how they make their living - the process should recognize these different roles
- facilitation by third parties or participants themselves (ie., humbling the brand of government) - the challenge is to establish distance between the government and the participants
- an expectation of perpetuity, but also an eventual hand-off to an NGO

From open source software communities (a "do-ocracy"), we get the ideas that:
- ways to make consensus decisions while maintaining place
- ways to effectively re-use prior work
- peer ownership and stewardship - we're all in this together, it's not just government asking for your opinion, but rather we're all doing something
- a reinforcing of open standards, and vice-versa
- a connection to reality - this is not an ivory tower process - code used in production serves as a grounding purpose

(Comment on how transparency is not an unalloyed good. But... there's a skill set. There's a way to make decisions, a way to have discussions, even when people are watching. Ways to have consensus even when not everybody has responded.)

Asked how to help, the 'Connect project' at HHS. (See O'Reilly article here.) A lot of the money in health care is spent by governments. Eg. wounded soldiers (just one example). They may touch a variety of different services. This group was charged with allowing patients to go from one to another and to bring their own records with them. This is the big problem with health care. The mandate was, 'come up with a standard to do this.'

The first effort was a top-down effort, to get the agencies to have 'must have' features in the standard. Eight separate consultants created these lists. A second try was to try something different, that would aggregate the demands of different agencies, and run them as pilot projects.

"CONNECT is an open source software solution that supports health information exchange – both locally and at the national level. CONNECT uses Nationwide Health Information Network standards and governance to make sure that health information exchanges are compatible with other exchanges being set up throughout the country." Apache (BSD) licensed. It's basically a gateway that can talk to other similarly-configured service. We created a 'partners list'.

(Is this a model for other countries, not just the U.S.? Nobody else has really come up with a systematic way of building this. A lot of the tech uses US-specific terms, data sets, etc.)

(Is stuff like this being mission-critical for businesses? The whole social media thing generally? More adoption of standards is important, it can help create good positive outcomes, using public collaboration as the tool to get there faster.)

(This is an interoperability initiative at core. But all initiatives that try to create interoperability by publishing large stacks of documents fail. But with open source, they are almost by definition interoperable. So governments that don't implement using open source - forget the licence, it's the source code there - are going to fail. It really feels that where government impacts people really happens day-to-day in the agencies.)

(Open health tools initiative - intended to be source that is released - they are building good libraries for parsing various health care documents.)