The information below concerns the UNG 1.0 project, for current efforts, see the UNG 2.0 project.

Project Overview

The UNG project intends to provide a complete alternative to proprietary cloud computing solutions, starting from enterprise applications and ending with search engines. The current development focus is Web Office and Web Groupware. We invite all Javascript developers to help the UNG project by improving community projects which have been select for UNG.

screenshot of the first prototype of UNG Docs

Status and Roadmap

The table bellow provides an overview of the status of UNG.

• Ready means that it is already used by mission critical projects
• Prototype means that it is already used by a few people
• Proof of concept means that a demo what produced based on which development can start
• Design concept means that ideas for a proof of concept exist

UNG Docs

The UNG Docs project has been officially presented during the Open World Forum in Paris, in November 22^th. The Prototype is made of two Business templates which are downloadable at:

• svn.erp5.org/erp5/trunk/bt5/erp5_web_ung_core/
• svn.erp5.org/erp5/trunk/bt5/erp5_web_ung_theme/
• svn.erp5.org/erp5/trunk/bt5/erp5_web_ung_role/ - Optional

Releases:

• 0.1 (2010-11-22)
• 0.2 (2011-01-19)
• 0.3 (2011-02-04)
• UNG Docs 1.0 - Press Release (2011-05-11)

Instructions to install them over ERP5 will come soon. 

UNG 1.0 uses ERP5, UNG 2.0 will be pure javascript with any backend.

After install the ERP5, you can use the configurator to install UNG in few steps

• svn.erp5.org/erp5/trunk/bt5/erp5_configurator_ung

Screenshots of UNG Docs and UNG Calendar:

• www.freecloudalliance.org/ung-Web.Office.Screenshots

Web Office Components

The current focus is Web Office and Javascript (JS). We will first try to leverage community projects and contribute to them. Only if this approach fails, we will eventually consider redeveloping components from scratch.

The key design principles are the following:

• develop pure JS components (ie. they should not depend on a PHP, python or Java backend) which can eventually be used with poor HTTP connexion and offline.
• develop framework independent JS components (ie. it should be possible to integrate them in any framework such as qooxdoo, sencha
• use W3C standards such as HTML or SVG (rather than ODF or OpenXML) for data formats.
• implement unhosted web standard presented by the myfavouritesandwich web application, and thus, ensure data security and privacy.
• contribute to community components until they can match the usability and features provided by Google Docs Microsoft Live

All components will then be integrated into TioLive platform which will serve as a reference platform and provide reference UNG Docs service on top of NEO distributed transactional database. All components should be designed to be usable with other platforms such as EyeOS, Plone, Drupal, etc.

Just like OpenOffice is currently independent of the operating system and of the file system, UNG Docs JS components should be independent the backend storage and application server.

The components which have been selected are:

• JQuery.sheet: a simple spreadsheet which uses HTML tables as data format.
• svg-edit: a simple drawing tool which uses SVG as data format.
• xinha: a simple text editor which uses XHTML as data format, supports XML extensions, produces clean code and handled semantic styling in addition to WYSIWYG.

Javascript Input/Output (JIO)

Our aim is to propose a new storage architecture. We want the owner of a file and the selected co-workers to be the only ones who can access it, and thus provide our users a real control on their data.

In our opinion, web applications mustn't store any data about users. It would ensure more privacy and more security. Moreover, it would decrease the cost of web applications which would no longer need lending storage servers.

The issue is to enable the application to access data easily. The unhosted web standard seems to already propose an interesting solution, and we think that we can go further. these are the elements of our architecture :

• an application server: it is the server who hosts web apps.
• an identity provider (webfinger): the role of this server is to provide ID information about users and, especially, the location of their jio.json file.
• the jio.json file: this file remembers the location on the cloud of each file of the user. The user's files are crypted with an asymmetric-key algorithm, and the private key is stored on this jio.json file, and encrypted with a user password.
• The cloud: the cloud provides an unlimited storage space for users. Each server connected to the cloud can provide storage space, and data privacy is ensured by the JIO encryption.

To summarize, The application server provides javascript methods to the browser. The browser asks the webfinger server and gets the location of the jio.json file. It creates a request and asks the user their password to get this file. Then, the browser can access the user's personal files and edit them with the javascript application.

The ambition of JIO is to provide the API needed to make all these elements able to communicate and implement this architecture.

Future

The next focus will be Web Mail and Web Calendar. For Web mail, a prototype application which is capable of handling 1,000,000+ size mailboxes exists. It supports full text search indexing of emails and attachments, with reponse time of less than 0.5 second to find 300 results out of 1,200,000 emails. For Web calendar, the implementation will derive from a CRM and support SyncML. The use of JQuery Full Calendar component is currently considered.

For the search engine, current ideas are kept confidential since they could provide a simple way to create an efficient distribute search engine for a vey small investment.