From CrisisCommons Wiki

Overview

This project page is for summarizing and tracking the development of a pict-o-speak mobile functionality under the Translation Applications project umbrella in coordination with the Languages and Translation project.

DRAFT - This project page has not been reviewed yet with all the key people working on the translation and mobile efforts, so its just a first cut at describing what we may/should do.

Some people have been working on this idea in Tradui, so please review that work.

Background

As described on Languages and Translation pict-o-speak books are used facilitate communication between people who don't speak a common language. In crisis situations there is often a large mix of languages and dialects coming together, pict-o-speak books can be used to quickly communicate key points visually across the various languages being used. They are also helpful when communicating with persons who are illiterate. Interestingly, these pict-o-speak books often may have little or no written text. They are usually navigated in top-down, category-based approach (e.g. find image for medical, then underneath the image for organ, then underneath the image for eye). This is quite different from how a written dictionary would be used (look up word eye and see translation "je").

Goals

Mobile Application

The project will focus on how to facilitate visual based communication using mobile devices (initially iPhone and Android devices). As described above, it will support navigating images to facilitate communication. Since it will be electronic, it can also support navigating to related terms. We also will be able to integrate with a written dictionary as well. Making it possible to show the written terms as well while navigating the images. The reverse is also possible, uses written terms to look up translations and images (however in practice the number of written terms greatly outweighs the number of images available).

Authoring Web Site

One of the key challenges is actually not the mobile application itself, but authoring the information to power the application. Pulling together a large set of images, labeling them, and then organizing them in a way that supports intuitive navigation to the images is a real challenge. This makes the authoring tool a key goal for this project. Hopefully we will be able to build an online authoring tool that will allow us to crowd source authoring, allowing people from around the world to submit public domain images and to help organize the navigation trees across those images.

Approach

Volunteers in Denver/Boulder as well as the Bay Area have done some work on drafting preliminary taxonomies for the categories below. These will be uploaded as they are completed. The next step would be for the group in DC to pick these up and use them in a first round of development.

More to be described here after the growing mobile team meets at CrisisCamp DC and pulls together a plan. Obviously we will continue to build on the accomplishments already achieved by the translation (English/Creole translation XML) and mobile (Tradui) teams. Exactly how we will manage authoring across the various translation efforts and the best way to turn that content in mobile apps and written documents still needs to be mapped out. Its likely we will end up with multiple tools, all hopefully sharing the same repository.

Taxonomic Categories

Taxonomies are being created in Freemind and are available for download from the CrisisCamp Lang/Trans Google Group.

Click category links to see Creative Commons licensed images that have been gathered for each category (taken from: OCHA Clusters Goals):

Health : To prevent excess morbidity and mortality and improve the health status of the affected population. The cluster forms working groups on issues related to communicable disease; reproductive health; hospital; medical services, supplies and rehabilitation; mental health and health promotion.
Emergency Shelter Coordination: To complement Haitian Government and civil society efforts in providing all earthquake-affected people with adequate, appropriate, and habitable emergency shelter materials in safety and dignity
Water and Sanitation: To prevent of water and sanitation-related diseases as well as enhancing the well-being and dignity of the affected population by insuring availability of water and basic sanitation facilities.
Protection: To ensure the protection of earthquake-affected people and strengthen local capacities to protect that population. The cluster forms working groups concentrating on issues, such as child protection; psychosocial; gender and disabled people.
Early Recovery : To lay the foundations for a sustainable recover process by supporting planning, coordination, monitoring, and business; rehabilitation planning for key cultural heritage sites and incorporating disaster risk reduction into overall recovery effort.
Food and Nutrition : To ensure that food and nutritional requirements of the most needy and vulnerable people are supported to facilitate their recovery. .mm file .pdf
Education : To support the Education Departments of affected provinces in providing immediate and medium-term assistance to re-establish child friendly temporary learning spaces.
Agriculture : To recover the food security and livelihoods of farming communities in the earthquake-affected districts with an aim to reduce their overall dependency on external food aid..mm file

Foto Pale (“Photo Speak”) Mobile Application Prototype

Foto Pale (“Photo Speak” in keeping the naming approach of Tradui) is a prototype of what a Pict-to-speak mobile functionality could look like. It allows a user to quickly navigate from a top-level set of images to more detailed concepts. It uses images that were pulled from the images made available on the translation team’s page. It also shows the English and Creole (when available) terms underneath the images by reusing Tradui translation engine.

To support crowd sourcing its content, Foto Pale actually uses a HTML page as its source document for it hierarchy. The concept is that members of the translation team (and others) can just edit the wiki page normally using bulleted lists to show the hierarchy. That wiki page can then be used directly (or with limited machine translation) by the tool. This technique may allow us to pull a basic hierarchy together quickly, but it’s not expected to scale as we add a lot more images or more complicate navigation. An authoring tool will be needed that supports image collection and concurrent users.

The current version of Foto Pale has very limited hierarchy of images and is most certainly a prototype. It does however demonstrate much of the functionality that a basic pict-o-speak app would need. It runs on the web, Android, and iPhone. It intentionally uses the same underlying frameworks as Tradui (Applicator Titanium, JQuery) with goal of making refactoring functionality across the two apps straightforward. The source code will be uploaded to GitHub at http://github.com/reesd/FotoPale- for reference and reuse sometime this Saturday (1/23).