Note: Edgar is discontinued in favor of Vix.

Welcome

Edgar adds portability and interoperability to web applications. It is an implementation of the Atom Publishing Protocol and offers a document publishing and management system. Documents are entries which can be grouped together in a feed. The content that can be stored goes well beyond flat text. The Atom Protocol supports media resources, which means that media (like images or videos) can be added to entries. Through the Atom Protocol, Edgar provides a flexible API to all data.

Edgar is the core implementation of this abstract foundation on top of which end-user applications can be built. These separate applications rely on Edgar to deal with the data, but provide a front-end that displays it to the user. Typical examples of these front-end applications are weblogs or image galleries. The benefit of using Edgar is that the system provides an API out-of-the-box instead of as an afterthought. With portability of data being a key feature of the new breed of web applications this is an important element in modern web application development.

Components

Edgar uses a lot third party components. The most important of these is the Pylons framework, but the application also uses SQLAlchemy, lxml and the Sphinx documentation system. The current focus is getting the version 0.1 prototype done. When this prototype is finished front-end applications will be developed, starting with a simple weblog application.

Roadmap

Edgar is still in the prototyping phase. This is a general outline of the planned features:

• version 0.1:
  • basic implementation of the Atom Protocol,
  • support for managing and displaying text-only entries,
  • basic authorization/authentication.
• version 0.2
  • support for media resources (binary enclosures).
• version 0.3
  • Amazon AWS/S3/CloudFront support for media resources,
  • ability to push data out to third-party services (like Blogger, Youtube, etc).
• version 0.4
  • scalability,
    • supporting data replication between mirrors of the application,
    • possibly adding support for distributed databases.
  • optimizing performance, including,
    • caching feeds (including a gzipped version to save bandwidth),
    • storing read-only "frontend" feeds (e.g. a weblog feed) on Amazon S3/CloudFront.
• version 0.5
  • search,
    • implement full-text search on documents using Sphinx, Lucene or Xapian.

Edgar is a brand new project, but the philosophy and concept is partly inspired on some ideas implemented in the abandoned nCMS project, which was intended to be a CMS application based on Pylons with an Atom Protocol API. Edgar, however, is fundamentally different from this project in both code, approach and design.

This revision is from 2009-12-01 16:52