Kaleidoscope :: Platform

Information overload is an everyday part of life-- inboxes overflowing with e-mail, hundreds of RSS and Twitter feeds, countless contacts across social networks, cell phones, and e-mail address books, etc. There is an urgent need to make sense of all this data to give it structure and meaning. Kaleidoscope is your cloud of knowledge.

Kaleidoscope is a free-form, Unicode-compliant, semantic database that enables a user to store data from various sources. Once in the database, it can be structured, annotated, and linked in any way imaginable. Data can then be among friends or with other web services. Interested in trying out a part of Kaleidoscope? Click here to learn more about Twitterbase. For information on the Kaleidoscope Twitterbase birds of a feather session at Web 2.0 Expo 2008 in New York, click here.

Below is a list of features:

* Flexible data store. The datastore was written to support Oracle BerkeleyDB. However, the API is written in a way to allow any datastore to sit underneath the Kaleidoscope API (think of this as a meta-database). Currently, most data processing functionality and limited support for triggers is currently available. Transaction processing is currently tied to Oracle BerkeleyDB but will be abstracted to support other datastores. Future datastore support could include the following: standard SQL databases, Amazon S3, Hadoop HBase, Facebook Data API, Google Spreadsheets, and more.

* Flexible data creation and management. Most existing database systems require the user to know in advance how the data should be structured and how data interconnects among each other. Kaleidoscope supports a model that could be called "entity-tagging" (an example of this is Metaweb Freebase). One could imagine an entity as a container for tags. Then, add tags with values to the entity. For example, I could create an entity named "chicago" and and give it tags of population = 2800000, state=illinois, timezone=utc-6. This style of data storage allows for flexible querying and viewing of data. For example, I could view all of the entities that have a tag of population or I could view all of the attributes for the entity of "chicago". Tags and entities can be deleted without causing a major rethinking of data structures.

* Simple text command interface. A simple text interface allows a user to perform the essentials of creating entities, attributes, and querying.

* Storage options. Kaleidoscope supports in-memory and disk-based data storage. The primary semantic data store is disk-based, but in-memory data stores can leverage this existing data store to provide integration with existing semantic structures on a more interactive level.

* Open source. The core parts of Kaleidoscope will be distributed into open source. This will allow the Python module containing Kaleidoscope to be used in countless applications.

Although the Kaleidoscope platform is in an alpha-level release, future features will include the following:

* Scripting languge. The simple command line interface will expand into a small scripting language which will allow simple data sharing across various communication models (instant messenger, Twitter, e-mail) and integration into existing web services and local applications.

* Communication options. Currently, Kaleidoscope supports Twitter for a user community. However, future versions will include any instant messenger protocol including AIM, Yahoo, and Jabber/XMPP. Through supporting these instant messaging protocols, it will be possible to have a computing device without a static IP address and be able to serve data to anyone on the Internet.

* Cloud computing and web services. Through the support of database triggers, an enhanced scripting language, and improved communication protocols, computing services can occur either locally or remotely. By mapping data structures across machines and with the use of XML-RPC and JSON-RPC, it will be possible to connect services in a simple manner. Through the use of Python ctypes, it will be possible to also integrate existing shared objects (like Windows DLLs or UNIX shared objects) into the mix to create next generation applications that support integration into both local and remote datastores. Data migration to and from remote data stores will also be possible.

* Improved security. Fine-grained permissions and group access control will allow users to determine what users can access. Additional security measures will include datastore encryption and prevent data tampering through the use of data checksum monitoring.

* Data recovery. Recover data from a previous time or undelete data that was destroyed up to a certain point. Changes to data will be recorded for a certain time frame. Deletions will be permitted to go to a container that behaves like a "desktop recycling bin". When a user is ready to destroy the data, simply empty this container.

* User interface viewer and web widgets constructor. Explore data using a traditional folders and icons interface but be able to construct web widgets using a graphical user interface builder. These widgets will run both locally and remotely and interact with the Kaleidoscope data store. Eventually these web widgets will be integrated into a new type of user interaction metaphor called Chamber.

Learn more about Kaleidoscope and general information about the semantic web through viewing these Powerpoint slides describing Kaleidoscope: overview, prototype of Kaleidoscope desktop interface (Chamber).

Imagine uses a database without the complexity. Imagine the possibilities of integrating data sources and services never connected before. Imagine having access to data at any time. Kaleidoscope empowers you to imagine the semantic web on your terms. Kaleidoscope allows you to process data your way.

TRY TWITTERBASE TODAY
If you have a Twitter account, follow the user kscope. Then, you can send private commands to Kaleidoscope that will allow you to store, retrieve and query your own private and public data as well as the public data of others. Remember, since Twitter is accessible through both cell phone text messaging and a web interface, you can have your data on the go when you need to know. When the demo is ready, you will be able to click here for a command list.

WEB 2.0 EXPO 2008 NEW YORK
Learn about the session here. View the slides from the session here.