Just another WordPress.com site

CrowdCam: Software Proposal

Here’s my final project of the Knight-Mozilla learning lab: a video pitch and software proposal for CrowdCam.

Watch the CrowdCam video pitch

Download the CrowdCam_software_proposal (PDF)

I look forward to all of your feedback!

Gathering Citizen Content: Pull vs. Push

Assignment 3 for the Knight-Mozzilla Learning Lab

Last week Mohamed Nanabhay (Head of Online at Al Jazeera English) wisely observed that news organizations are “no longer the sole arbiters of what people see and how they are represented.”  

He recalled that while covering the Egyptian uprising on January 23rd, the network was flooded with “citizen media” that it didn’t produce, often raw footage with natural sound shot on cellphones by protesters or bystanders. He said that Al Jazeera’s news gathering strategy is evolving to increasingly “go out and find citizen content.”

So, as a News Lab member inquired: as news organizations place increasing emphasis on citizen content, is it better to have the pubic send in material or to pull material from sites such as Youtube?

This question is central to the justification for my project CrowdCam, and so I’ll reflect on two different news gathering approaches for citizen media, which I dub Pull vs. Push.


In this approach, embraced by Al Jazeera, news organizations curate existing “citizen content” by pulling it from websites. The footage is free; avoids ethical and legal pitfalls of putting citizens in harm’s way; and taps into existing citizen practices and technologies. Yet the footage lacks exclusivity (i.e. anyone can get the same video off Youtube); there are vast amounts of material to verify and curate; and this passive approach is quite tedious and slow.   Here’s a clip from Reuters, posted to Youtube.

Note the curious feedback loop: Content is posted by citizens to Youtube, compiled and curated by Reuters, then reposted to Youtube (another feedback loop is poetically reflected in the photo: a cellphone camera photographs another cellphone camera).  Amateurs can remix and edit their own compilations, and the results are often far more popular than the efforts by news orgs.  This amateur video has received 2M+ million views while the Reuters compilation has only 28K views.


In contrast to the above approach is CNN’s iReport, which pushes out specific calls for citizen “assignments.”  Then CNN posts what they get back.  Using this method, how can CNN guarantee quality and journalistic integrity?

It’s not easy. Note how CNN nervously places this pop-up disclaimer on every page (until disabled by the user): “The stories in this section are not edited, fact-checked, or screened before they post.”

To its credit, CNN employs a reputation system to help viewers vet the credibility of the source. One user, “Bhatnewpics” (a self-described journalist, PR professional, social worker, and former farmer), has somehow had the time and motivation to upload 866 iReports.

The BBC has embraced a bit of a hybrid approach: pushing out requests and then pulling in citizen media.  There’s a call for submissions upload page backed by an impressive User Generated Content division, with more than 20 employees, to carefully vet the quality, provenance, and authenticity of the submissions.  Note the language: “If you have a news photograph or video that you have taken” (my emphasis).  I may be reading to much into the passive voice, but it seems an intentional effort to not actively solicit “assignments” that may put citizens in harm’s way and to reduce the BBC’s ethical and legal liability.

Notably, the BBC includes this disclaimer:

This points to a cautiousness on behalf of news organizations to become more directly involved with citizen newsgathering.  Mohamed Nanabhay also seems wary of the approach, saying “you can’t expect people to send you good content.”


There’s no doubt the above Pull and Push approaches can be useful, but they’re both limited: slow paced during fast-moving events, requiring intensive resources to vet and curate citizen media, and there’s little guarantee of quality.

My project, CrowdCam, proposes a bold new approach: a request-based system that allows a news organization to get exactly what it wants, fast.  This is possible through the power of live 2-way links between newsrooms and citizens via video enabled mobile phones, enhanced by geolocation, a marketplace, and a strong reputation layer to encourage transparency and quality.  There are challenges — such as establishing a community with wide geographic reach, tapping into appropriate technology, and encouraging new user behaviors — but I believe such a transformative system is inevitable.

Web Apps to the Rescue?

Assignment 2 for the Knight-Mozzilla Learning Lab

Last week’s lectures by Christian Heilmann and John Resig sparked a bit of a revelation about my project CrowdCam: I now realize that CrowdCam might reach a much larger audience if it could be implemented in HTML5JQuery, and CSS3 rather than as an iPhone app.  This could help solve a major piece of my new problem statement: “how to find people anywhere in the world, fast?”

Until now, my ideal scenario was that millions of people would rush to download a CrowdCam app for iPhone, Android etc.  The app would include a rich feature set and tight integration with the phones’ camera, GPS, and other hardware.  The app could chime with alerts, while using other apps or asleep in a pocket.  It would do everything.  But in the real world, it’s pretty hard to get millions of people to install an app and keep it updated over the years — especially if they rarely get pinged with requests for things to film.

But what if there was an easier way to disseminate the platform?  HTML5 and libraries like JQuery may point the way forward.  Dynamic touch interfaces and rich user experiences are no longer confined to iOS.  See the Web app for Financial Times.

Best of all, Web apps require no download or user updates.  So I could theoretically distribute plain old links that would open up a Web app with all the features I need.  Today, slick interfaces and geolocation are possible.  Yet there are still some unanswered questions, in particular: can I tap into (or hijack) the phone’s camera and live stream from a browser?  That’s key for my system to work.   But if I can cobble that together, and creatively use other existing and simple technologies (text messages instead of push alerts?), I may be able to replicate most of the iOS app feature set in a package that is cross-platform, open source, and accessible with a single click of a URL.  Wow.

After our Wednesday lecture, I was able to chat with one of Mozilla’s technology gurus, and he has given me a glimmer of hope that is may all be possible… the question is when, and what to build until that day arrives.

Solving the Right Problem

Assignment 1 for the Knight-Mozzilla Learning Lab

I’ve been thinking a lot about this bit of wisdom from Aza Raskin’s lecture: Most of design is finding the right problem to solve.  He recounted the story of Paul MacCready, who won a longstanding airplane engineering competition by reframing the design problem: he shifted his focus from building a complex machine to building one that could be repaired quickly.  As Aza later summarized, “It’s not about thinking outside the box, but finding the right box to think inside.”

This resonates strongly for me and my project, CrowdCam — a crowd-sourced mobile platform to provide live video news feeds from virtually anywhere, at any time.  In recent months, I’ve begun building basic interactive GUI mock-ups of the full system (an iPhone app and Website).  I’ve labored intensely over every screen, menu, and icon.  Adding hotlinks, I’ve simulated the interaction flow.  Though rough, it’s pretty exciting, and I can’t wait to work with engineers on building the full prototype.

mock-up screen for CrowdCam app

At the same time, I’ve been seeking feedback from media executives, cyberlaw scholars, civic media mavens, and technology entrepreneurs.  The recent MIT Knight Civic Media Conference was especially helpful.  People generally love the idea.  There seems to be a real need and niche for a system such as CrowdCam.  Yet I’ve also been hearing a consistent line of questioning that isn’t so easy to answer.  How large a network of registered “Cams” (people with video-enabled cellphone cameras) do I need? And how will I sign up all these people?  Will people in remote areas, even war zones, have the right technology?  If so, what would prompt them to download the app and register?

These questions underline a troubling possibility: What if I build the most beautiful, elegant system — yet not enough people use it?  Without scale, the platform will be useless.  It’s one thing to build a technology; it’s quite another thing to build a community of users.

So now I’m wondering if my principal design problem is not: “build the best platform for mobile live streaming and 2-way communication to newsrooms” (even though that is of course a key component).  Like Paul McReady, I may need to shift focus from building a complex machine to solving an underlying process.  The new problem statement might be: “how to find people anywhere in the world, fast.”  Can I use existing technologies?  Piggyback onto existing communities?

Looking over Aza’s slide of his top 7 design tips, #5 seems right on: “You will change the problem you are trying to solve.”  Though as I click through my proud mock-ups, I’m not quite ready to embrace #7: “Plan to throw it away.”