Agile Author Overview

Introduction

Agile Author is a framework for developing networked repositories of digital information such as digital libraries and content management systems. Agile Author (AA) represents a unique approach to the storage and retrieval of digital information and is appropriate for a wide range of needs. This document will only highlight the most prominent and commonly used features of AA and will not be able to directly address every potential question or need. But the strength of AA is that it is powerful and flexible enough to be used in projects whose needs can't be easily foreseen or whose needs may change dramatically throughout the life of the project, which includes most projects.

The following are some key aspects of AA:

• AA lowers the software, hardware and knowledge barriers to creating and updating content yet offers many advanced features.
• AA allows for rapid prototyping that enhances the quality of communication between developers and clients, yet it is also a robust and scalable enterprise solution for large, long-term projects.
• AA has many built-in capabilities, yet it also makes few assumptions about how it will be implemented, allowing for more flexibility.
  
• AA allows for arbitrary organization of information, yet the resulting information is easily navigated, searched and understood.
  
• AA is based on reusable components so that the code base for each implementation is very small and updates can automatically be made across implementations, yet it is easily customizable for a wide range of applications.
  

From the client's perspective, AA is a tool with which to create and update their own digital content in a manner that is both easy and flexible and which automatically creates presentational and navigational components appropriate to the published environment. Clients very much want this focused control of their information without being required to think about the details of the publication of that information in a particular medium, and they have responded enthusiastically to AA as a result .

In its most common use so far, university faculty (and often their graduate students too) use AA to create their own Web based resources, quickly building large, complex sites that feature automatically updated searching and navigation; they do so on their own time, and at their own pace, and with little or no assistance from designers or programmers. The dual effect is that the sites are used heavily both for content production by faculty and consumption by students, yet from a support perspective, we almost never need to assist the client after the initial deployment of AA despite its continued rate of use.

Example applications of AA

• As a digital repository of images (or video, sound, etc. or any combination of digital formats)...
• AA supports large objects and large repositories.
• Available image manipulation capabilities include digital watermarking, thumbnailing, and panning/zooming.
• Metadata can be configured flexibly, such as assigning subject headings retrieved from a real-time query of the library catalog.
• Advanced searching and browsing options are available, including features such as user bookmarks and user-defined categories.
• For Web-based course management, distributed learning support, and online tutorials...
• Create courses with original content and materials supporting lectures.
• Add interactive quizzes that can be graded according to a custom, pluggable evaluation method and customized with feedback and navigation options based on the outcome of that evaluation method.
  
• Create reusable "learning objects" directly from existing content.
  
• Allow students to engage in threaded discussions of published materials.
• For "Virtual field trips"...
• Manage subject/context based information organized by location (cities, facilities, etc.) and aspects of each location (food, weather, laws, policies, etc.).
• Display groupings of multimedia information for each aspect of each location.
  
• Dynamically cross-cut information, such as comparisons of weather across cities.
  
• As a bridge between online and print publishing, such as a pre-print server...
  
• Multiple authors collaborate on information using AA.
  
• Information can be published via the Web and reviews/comments can be attached to each object or entire sections.
• Information is rendered in a standard XML format that can be reformatted for Word, PDF, DocBook and sent to a publisher electronically.
• For digital asset management and peer-to-peer sharing of data...
• Quickly create a desktop based client to AA using an XML based user interface language (XUL).
• Use OAI or similar resource indexing and discovery mechanism to "publish" information to other authenticated users of the XUL based AA client for review and collaboration.
  
• As a "PowerPoint for the Web"...
• Create simple slide shows that become a flexible, Web-based alternative to PowerPoint.
• For simple, organizational Web sites
• Author and update a simple Web site that is automatically searchable.
• Use the built-in one-click workflow to build large portions of the site privately, then easily publish whole sections with one click.
• The site's navigation automatically reflects the published structure of the information.

Available interfaces

• Default Web-based interface:
• XHTML 1.1 compliant.
• 100% CSS driven.
• Exceeds Federal Section 508 and WAI AAA priority guidelines.
• Templates codify common UI patterns for information display and navigation.
• Automatically generated navigation includes an expandable tree, breadcrumb trail, Google style paging, and 'slide-show' navigation.
• AA detects and supports compressed communication with browsers.
• Edit text objects in a WYSIWYG editor that includes advanced features such as cleaning pasted Microsoft Word text.
• Automatic thumbnailing and resizing of images is provided. These versions of the image are cached for performance but updated whenever the original image data is updated.
  
• Split public view/private edit modes provide immediate feedback of the effect of every change made.
• A 'compact' editing view is provided to better accommodate large amounts of information.
  
• Printer-friendly views of the information can be generated automatically for single pages or entire sections.
  
• A 'glossary' view automatically gathers all defined term objects scattered throughout the repository, sorts them (accounting for initial articles) and displays them as a glossary with alphabetical navigation.
• AA provides integrated user management with complete security infrastructure of roles and permissions that can be specified at various levels of granularity in the repository.
  
• Login screens are automatically invoked if information is protected. Options exists to allow users to self register and reset lost passwords.
• Complete Open Archives repository, harvester and aggregator components are available including the ability to allow more advanced searching options than are typically found in OAI repositories.
  
• Complete FTP and WebDAV support is available.
• Other XML based protocols are supported either directly (RSS) or via XSLT (DocBook).
  
• An arbitrary number of interfaces can be created for the same data independently of the data itself.
  

Data storage

• Object types supported include:
• pages,
• text,
• images,
• files (including sounds, video, and documents such as PDF, Word, etc.),
• hypertext links,
• questions,
• answers,
• defined terms,
• annotations, and
  
• cross references.
• By design, every object is a "folder" able to contain other objects; data is hierarchically organized and objects are aware of the hierarchy.
• Objects are also related in circular and cross-cutting ways, such as by type and through cross references.
• By default, objects are stored in an object oriented database.
• Storage is pluggable and can be made to support any storage mechanism. Currently, AA objects can be stored in a relational database (many major vendors supported) or the server's filesystem.
• Every object is automatically indexed for searching and logs every user that interacts with it.
• Every object has a stable URL.

Metadata

• By default, Dublin Core metadata exists for every object.
• Any XML format can be supported either directly or via XSLT. Several formats are already implemented.
  
• Metadata is extensible via simple configuration files to easily support arbitrary metadata standards.
• Metadata can be uniformly specified across all object types, based on individual object types or some other criteria.
  
• Data such as controlled vocabularies can be easily retrieved from any networked resource allowing for simple integration with existing systems.
  

Searching

• Every published object in the database is automatically indexed for searching using widely accepted information retrieval standards including:
• the ability to define multiple indexes of various types such as field, date/time, keyword and full-text, and
  
• the retrieval of lightweight "metadata" objects as surrogates of the actual objects, while allowing easy access to the objects represented.
• Standard search engine features are supported, including wild cards, boolean terms, proximity searching, synonyms, stop words, left truncation, stemming, and relevance ranking among others.
  
• Indexes are configured via simple text files.

Built-in user management and security

• By default, published information is publicly accessible, but it can be protected with a fully featured set of permissions that can be specified at various levels of granularity.
• By default, three roles exist: anonymous, authenticated, and manager, but new roles can be added.
• Random, complex passwords can be automatically generated and sent to users.
• Depending on configuration, users can self-register and reset lost passwords.
• Those in the "manager" role can also register users and reset passwords.

Enterprise features

• AA is based on a standard "LAMP" architecture.
• AA is built using accepted design patterns and best practices, most obviously, Model-View-Controller.
• AA is cross-platform: current deployments on Solaris, Windows, Mac, BSD, Linux, HPUX.
• AA can run as a stand alone server or be distributed across multiple servers.
• AA can be integrated into many different architectures through pluggable front and back ends.
• Object persistence and indexing is transactional.
• Objects, displays, searches, etc. are easily cacheable.
• AA and its components are 100% open source software with commercial hosting and support available

Future development

• Make workflow configurable in a similar way to the metadata (partially implemented).
• Allow editing and viewing of mathematical equations without special browsers or plugins (partially implemented)
• Incorporate Zoomify (partially implemented).
• Build in support for exporting to PDF, DocBook.
• Allow exporting to static/offline files (partially implemented).
• Provide indexing adapters for Word, PDF, OpenOffice (partially implemented).
  
• Build in awareness of segments of sound, video files.
  
• Retrieve/index/store "abstracts" of data represented by hypertext links (partially implemented).
• Provide global search and replace for text objects.
• Make undo/redo/history capabilities available to default Web interface (partially implemented).
• Provide better support for internationalization (I18N) and localization (L10N).

You can download Agile Author from the Agile Author project page on SourceForge.net.

Commercial development, hosting and support for Agile Author are available from Agile Enterprise Software, info@agile-software.com.