Distributed Media Annotation (DiMA)
After 30 years, WWW is entering the new world of low-literates who have typically managed information through storytelling. DiMA framework and Indigo project is about creating distributed archival and annotation capabilities in local community spaces. Indigo tools help with federation, media annotations and story telling of typically text-heavy Web content.
note: Internet has until lately been available to low-literates in a service driven model and not in the way information is available to the literate on the Web. This service driven model where services such as banking operations and provisioning of market prices for the farmers are sub-optimal as they do not bring the Web to them as we (literates) use the Web.
A media annotation framework needs to be deployed in both always-online and typically-offline networks. The framework has to support management of archives and annotations by a number of groups associated with the content in the archives. As an example, lets imagine that a number of groups are generating audio, video and text content. These are available as feeds that can be harvested, curated and repurposed by another service. Annotation on the audio, video and text material will not only help in categorizing and navigating but also in curating the content by the service. Annotations can be private, local and sharable. They also help establish deep links in the content which can help in both automating navigation to related content and to assist in developing multi-media storytelling by the communities. The federated data can be further made available for others as feeds.
Such a framework would work on standard protocols, provide common tools for annotations, involve feeds management and encourage storytelling services. In addition to this, the deployment of the framework would need to be made simple, people oriented and sufficiently documented. It would also use most promising technologies with respect of ease of deployment and ease of use - browser users and smartphone users.
Typically, the framework will include ID management, content and annotation management, media annotation tools and applications that work on common browsers and smartphones. These would come with Mozilla browser plugin, Pocket extension that helps bookmark deeplinks (annotation bodies/targets) and tutorial applications in local languages.
The renarration architecture builds on the available annotation services. It can be constructed as client libraries, annotation repositories, additional third party services and browser extension plugins. Some of the desired third party services are:
i) Finding trustable people as in social networking and forming groups for collective annotation-workflow needs,
ii) Workflow managing services such as in the case of issue management or a special effect rendering (say as A/V) of content based on a query that can be constructed based on the context of the domain of data,
iii) Automated and semi automated services such as translation or spell checker or machine learning based image matching as in the case of face tagging, and
iv) A dashboard for managing and making sense of the collections of the increasing linked-data across the repositories of interest.
Meta-services, such as handling spam and notifying abuse indicators may also be required.
1 Documentation of how one can annotate, curate and storify by oneself
ID management (phone number based/mobile first) Database/repository management Dashboard for distributed annotation viewing/management Media Annotation Tools Making it easy to deploy on local community servers
3 Planned Mozilla integration
Browser plugin Pocket extension to bookmark deeplinks (annotation bodies/targets) Private Public collaboration by interlinking annotations Also https://goggles.mozilla.org/ And https://whatcanidoformozilla.org/ as a way of introducing, like diff moz projects