Thanks for the warm welcome and thoughtful comments!

My optimistic prediction about the day when we have a paperless field is indeed wishful thinking. I hope I do live to see that day. But my prediction was in no way intended to suggest that we are anywhere close to it being ready to leave paper behind. Talking Papers, if it performs as described, will be a clear affirmation that paper has many years, if not decades, of service yet to render to the humanitarian relief and development communities. Given the natural advantages that paper has over digital formats, I don’t believe we’ll see a truly viable successor until we have a cheap, flexible, easy to reproduce “digital paper” technology that combines the best of both worlds. Although I am certainly excited about recent advances in mobile data collection (you may recall that I chair the Open Mobile Consortium), I’ve weathered the same seemingly endless progression of clunky, unsustainable, unscalable, proprietary data collection + map apps running on ruggedized PDAs that you have. Paper has a long life ahead. My hope is that Talking Papers might help to streamline how paper in the field is used to feed centralized information management systems on the back end, where its clear that technology is superior in so many ways to paper-based processes.

I can’t tell from your post whether you were believed that there was a commercial enterprise at work. As far as I know, no one involved in Talking Papers has discussed commercialization of this project. We all believe in open source, open standards, and open innovation. We want this project to come together as a community effort. Although free software is of course like the free goldfish at the pet store – the care and feeding is anything but free – we are all quite keen to make sure that what results is available to everyone who wants it for free.

I also agree with Michael on the need for greater supply chain visibility, and its quite possible that significant progress in that domain would improve the effectiveness of humanitarian operations more dramatically than Talking Papers ever could. That said, I do believe that Talking Papers could go a long way toward improving data collection, and that such improvements would also yield tangible benefits to emergency response.

You state that

“The basic fields required to respond to a humanitarian crisis are almost invariant no matter what the specifics of that crisis are. The main requirement is to take into account local technologies (e.g. do we need to look for boreholes or surface water supplies), terminologies (e.g. what words do people use to describe their situation) and ontologies (e.g. what are the administrative boundaries).”

Well, sort of. It is rather troubling that we as a community still can’t reach agreement on a manageable number of indicators to include in a humanitarian crisis Minimal Essential Data Set. I’m still traumatized by my experience working on the “Rapid” Assessment Form for Iraq back in ’03, which consisted of 338 data elements. MEDSs are a tangent that could take us down a pedantic rabbit hole and we’d best not pursue it here, but I think it’s worth remembering that data is collected during a response to help individual organizations fulfill their various mandates, and those mandates themselves evolve as coordination improves. I’d love to get the point that we could create forms on the fly that include means to collect data for both universally appropriate MEDS indicators and locally-appropriate indicators (it’s actually already slated as the topic of a near-future blog post), but the reality on the ground, as you know, is that that forms change all the time, often after vast numbers of paper copies have been passed out to people who can no longer be reached.

Regardless of whether the root causes of this constant flux are largely attributable to the factors in the political sphere, form designers, database administrators, and data entry staff bear the brunt of it. The self-describing data+schema payload of a Talking Papers form is intended, in part, to support scenarios where wave after wave of evolving datasets have to be compiled into a repository that has no prior knowledge of updates to schema. That is, Talking Papers can’t help prevent the flux. It might help field personnel accommodate it in some useful ways.

“Unless Robert has access to research that I don’t, my personal experience gives me no reason to believe that the levels of error in paper-based data collection are sufficiently high to significantly reduce the effectiveness of humanitarian coordination.”

My assertion was simply based on the recognition that 1) the inevitable misallocation of resources during emergencies is (in my personal experience) sometimes the result of misinformation, and that 2) technologies such as PDAs can perform client-side validation and catch errors at the point of capture in ways not possible with paper. I don’t have any hard numbers here on error rates (or lost forms) and its operational impact. That would be useful research indeed. My point was to suggest that there might be another step in the process – data entry – where a system like Talking Papers could help to improve another aspect of information flow: timeliness.

Finally, I’m not assuming that a staff to perform data scrubbing will simply materialize as the result of the availability of the data scrubbing tool within Talking Papers. Field staff and HQ staff are already scrubbing humanitarian data in spreadsheets and Access databases that was collected from paper forms. I’d like to create a tool that makes better use of their time by detection and visual flagging of potential areas in need of scrubbing and collaboration features for those who have access to the Internet (a luxury far more likely to be available at HQ). I’ve personally observed a humanitarian IM anti-pattern that I’ve selfishly named “Kirkpatrick’s Conjecture”, which is that the probability of a relief worker’s ending up manually scrubbing data all night in the corner of a tent is in direct proportion to their seniority within the response. Getting the data right before passing it up the chain matters so much that those who are already burdened with responsibility often don’t feel comfortable delegating the task to anyone else. Fixing that problem is going to be tough, but in the mean time, I’d love to put better data scrubbing tools in their hands to help them get at least a couple of hours of sleep.

There’s some wonderful work on the project already underway here. Please join the group and share your concerns, ideas for features, thoughts on prioritization, and any potential opportunities to test it in the field.

The Google group is here:

http://groups.google.com/group/talkingpapers

Mike Migurski’s Github repository is here:

http://www.talking-papers.org

As you emphasize, transparency is another weak point in this flow, and I’d put forward that it would contribute towards better use of information in decision making due to the increasing ease of accountability. Talking Papers should emphasize transparency and openness in its design as well.

While indirectly related, I thought I’d share this project which seeks to minimize error in data entry using mobile phones:

http://irevolution.wordpress.com/2009/04/22/improving-quality-of-data-collected-by-mobile-phones

I write “indirect” because most of the points you raise still apply to mobile phones. So the latter is no panacea but the project is still interesting and could have a positive impact in some cases.