This is the first of two articles about Mobile Voices, a project based in Southern California.

i-9e4714e78dafe8ed482c117fc53224ca-vozmob.jpg

Voces Móviles / Mobile Voices, a Los Angeles-based citizen media project, is a collaboration between the Annenberg School for Communication at the University of Southern California and the Institute of Popular Education of Southern California. In its own words, Mobile Voices is “a platform for immigrant workers in Los Angeles to create stories about their lives and communities directly from cell phones. VozMob helps people with limited computer access gain greater participation in the digital public sphere.”

For the last eighteen months, three programmers have been working on Mobile Voices. I talked to Sasha Constanza-Chock, one of the developers of the project, about the process that went into building the Mobile Voices platform. I asked why Mobile Voices built a custom platform for the project, why they chose to use multimedia messaging for content delivery, and the lessons they have learned about using MMS.

Goals

David Sasaki, an Idea Lab contributor, previously wrote a good introduction to Mobile Voices. The project’s primary goal is to develop a platform through which day laborers can broadcast their voices to a wide audience using mobile phones. The software needs to be scalable in order for it to be used by many day laborers. (The initial group only had 10 laborers, but a much larger audience is expected.) Another goal of the project, which is especially relevant to building the software, is the idea of using so-called participatory design: the technology would be designed according to needs and requests of the laborers, and built with their direct input.

The Platform

The current Mobile Voices platform is a customized Drupal system that lets day laborers blog by sending an MMS message to a Mobile Voices email address. Users can also text in via SMS, or call a local number to leave an audio message. They can register to start mobile blogging just by using their phone. They can also subscribe to blogs using MMS, translate posts, and more.

Why Custom Software?

Sasaki, in his review, offers three points of criticism (or “push-back” as he calls them) about the Mobile Voices project. I started my interview with Constanza-Chock with Sasaki’s third point, questioning the cost and resources needed to build a custom Drupal system in favor of using a pre-existing commercial software platform like Brightkite, for example.

Constanza-Chock noted that there were four reasons: participatory development, scalability, customizability, and privacy. To enable participatory design methods for the project, the Mobile Voices team wanted to be able to modify the software in any way requested by the users. They did not want limitations imposed by a commercial platform to hinder a particular feature the day laborers wanted.

Moreover, Mobile Voices needed targeted scalability because they eventually want to extend the platform so that any day laborer could use it. Day laborers also tend to use low-end phones, and most commercial platforms are geared toward tech-savvy smartphone users. (For example, iPhone apps are often favored over lower-end phones.) VozMob anticipated bending over backwards to fit each of its user’s needs, and needed a customizable platform.

Constanza-Chock also explained that an open source framework was chosen so they could protect their users’ privacy completely — a must when dealing with immigrant workers.

Why Choose Multimedia Messaging?

MMS is rarely used in mobile projects focused on social development. They typically use SMS, WAP browsing, the Internet, and even voice-based applications. However, the VozMob team uses MMS heavily. They even wrote software to combine a set of pictures, audio and photo captions to make up soundslides in blog posts (see an example here — click on the picture to see the slideshow). So I asked Constanza-Chock why VozMob chose to use MMS as one of its primary phone technologies.

VozMob, true to good participatory design principles, conducted extensive user surveys. The team examined which phones, plans, and types of non-mobile Internet access the laborers had. Seeing mostly voice-based usage (only 31 percent of respondents sent texts), the team proceeded to build an audio-blogging system accessible by simply calling a number. They used a simple voicemail-based system which they could access from Drupal through email, with minimal effort.

Soon, users were asking how they could upload pictures and other content to the site. The initial survey results showed 47 percent of the laborers take pictures, and AT&T prepaid plans (which a fair number of users were using) had cheap messaging plans. So VozMob decided that the technology that fit the users’ desires and needs best would be an MMS-based solution. (Constanza-Chock tells me they also considered J2ME-based applications, but the prototypes showed little promise).

How Has Using MMS Fared?

In the United States, all mobile carriers offer email gateways to their SMS and MMS services. The VozMob team chose these email gateways for getting their users’ MMSs. While Drupal already had a module that would import email, the team needed to put in some work to specifically receive email that came from MMS.

One basic difficulty working with MMS-generated email is that different carriers format messages differently. Some carriers attach text or multimedia ads for their networks, others send WAP links to media files (but no file in the message), and still others send files as email attachments. VozMob wrote the filtering and processing software necessary to get multimedia content in each of these different formats.

Another problem VozMob encountered was that unlocked phones are often hard to configure to use MMS. Constanza-Chock said providers were unwilling or incapable of working with the project to enable non-carrier-bought handsets to work with MMS, and highly technical engineers stumbled in making some phones work.

On the other hand, the team also encountered some pleasant surprises working with MMS. According to Constanza-Chock, most phones (even low-end ones) can compose elaborate MMS messages. With some digging, phones can compose a single MMS message that contains multiple pictures, sound, as well as text (the example from above was created using a single MMS). Mobile Voices has used this functionality to allow laborers to blog detailed slideshows, all using one MMS message.

What Worked Well

After a year and a half of development, Constanza-Chock is happy that most of the core functionality (for uploading) is there. Day laborers can upload photos, audio, and video using cheap MMS messaging. And they are excited about doing so: Constanza-Chock said users have introduced their friends to the system.

For the development process, Constanza-Chock said their use of an issue tracker was very helpful. The project ran participatory workshops to help design the software. The laborers would report problems and desires, which would turn into bug reports and feature requests on a RedMine issue tracker. Face-to-face code sprints, as well as face-to-face meetings between developers and the clients also helped the project along.

Challenges and Sticking Points

When I asked Constanza-Chock about the challenges the project faced, he pointed out that the telecommunications operators were hard to deal with. Not only did they refuse to work with MMS configurations on unlocked phones, they had unpredictable price changes in MMS packages. Constanza-Chock was also frustrated by expiration dates of MMS messages bought by clients. (AT&T prepaid sells messaging packages that offer a specific number of messages, but they  expire after one month).

Constanza-Chock also said working on a project such as this requires a lot of time and energy. There are often many possible ways of accomplishing a goal, especially when using free and open source software. Simply evaluating the pros and cons, and finding the right tools, takes a lot of work. A sustained project that is designed to provide a certain service to a user base requires sustained time and energy, Constanza-Chock said, which can be challenging.

Future Work

The project is not yet complete. Constanza-Chock told me a few ways they are looking to expand the project. The team focused on building the core software, and hasn’t spent a lot of time on marketing or outreach to day laborers. The Institute of Popular Education of Southern California, the resource center the laborers use, is building training kits so more laborers can start blogging on the VozMob platform.

The project will also focus on actually broadcasting the laborers’ voices. Going forward, the Mobile Voices blog will get a face-lift, there will be more publicity, and they plan to add tools to share and spread the workers’ stories using social media.

Finally, the VozMob developers will also re-factor the code they have produced so far, so it can be integrated into Drupal modules. The work on MMS and SMS filtering and processing using email gateways will be put back into the Drupal mailhandler and SMS Framework modules.

Part two of the Mobile Voices article will focus on the challenges of implementing a citizen media platform with marginalized populations. This series is cross-posted on MobileActive.org