At the moment, this page is more of an attempt to look back and summarize the past as a means of planning how to bring together the various outcomes of the project.
A short restatement of points from the Manifesto
- Encourage writeable archives
- Support a decentralized web / distributed / thrives on copies (mimiphilic?!) where copying and distribution strengthens the "usability" and richness of the archive
- Finding a balance in responding to current web and archiving practices in seeking solutions in open standards
- Meet each of these goals in producing useful software tools based open/standard/free protocols and tools, such as RDF, and concepts like linked open data
- Supporting workflows that allow multiple, distributed, contestable and ongoing outcomes
Distributed has technical implications, such as architectures that allow for multiple servers, or in any case avoid reliance on a single centralized one. Cloud computing in an interesting and related phenomenon in that different technical solutions (involved in distributed code) reveal a range of issues practical, financial, geographical/(inter)national, political.
A second aspect however is more editorial; many online content management systems, even radically participatory ones such as wikipedia, encourage a centralized and consent-seeking editorial process. How could a distributed CMS allow for distributed (many servers), localized (to specific contexts/institutions), and potentially contentious / discordant views coexist and strengthen each other rather than be competitive or unviewed through isolation.
Code libraries are condensations of practice; AA has been influenced by either direct use or study of several popular code libraries related to the web
AA has developed through a series of workshops and practical applications. See History
Software / Tools
Main software outcomes of the project.
Popup media player (in the tradition of lightbox/fancybox). Used in Toneelstof
Python script to render a dynamic site as a "portable" set of static files
- Shift from API/Django-specificity to a suite of tools that are joined by a consistent use of (RDF-defined) data models
- Django: Shift to new rdfindexed view decorator (instead of "indexed model" settings mess)
- Django: Move to more use of RDF, avoiding redundant Django Models
- The Core is:
- an RDF index
- getting media-specific metadata from URL
- a "meta" wiki or thesaurus/ flexible taxonomy tool where terms are managed, grouped as synonyms, filtered, remapped
- accessible via SPARQL & other (minimal) functional APIs
- provides filters as a means of transforming and cutting up resources
- provides powerful over views as tables, graphs
- supports various outputs as feeds...
- New ways of addressing resources (but online and offline), including:
- ... permalinks and "content-addressing" of resources (as opposed to URLs) / snapshots
- ... consistent support for fragments, as both a means of cutting up and addressing parts of resources
- Support for flexible embedding of media
- A "media player" that embraces all the media of the web (audio/video, documents)
- A Sniffer:
- Maps a URL to RDF
- Adapts existing APIs/tools to produce RDF describing online resources
- is decoupled from a view (through RDF)
- Media Player (aabox)
An abstract timeline as a means of cross linking different materials. Part of our approach is the use of Document fragments as annotation
- Downloading a video, watching the "online" version (with annotations) using your local copy (selectable which version is played).
- Same thing for annotations ?!
- Image embed: thumbnail, click to see "bigger", download link to original.
- Multi-source playlist (Toneelstof)
- Cross-institutional image gallery