Over forty years ago a man named Dr. Douglas C. Engelbart was working at Stanford Research International (SRI). The history is fascinating but I need to stay focused. You can read about the Augmentation Research Center, the 1968 demo of oN-Line System (NLS) held in San Francisco and Dynamic Knowledge Repositories (DKR). In 2001 I was fortunate enough to get a login and use the one NLS system still in existance (to my knowledge) in the pursuit of the next NLS/Augment system named the Open Hyperdocument System (OHS). The central idea behind the OHS Doug has formed into a document describing a Hyperscope.

What's a Hyperscope? It's a very very enhanced client through which information can be viewed and manipulated. Similar verbs include filtering, caching, transcoding and translating. Web based intermediaries and lenses are nouns that allow these verbs to be performed. A hyperscope can be seen as a lens that interprets and enhances data. It also can be configured (depending on the skill of the "knowledge worker") to display data in summarized form. There are many attributes to make a full Hyperscope.

It also has a "matching" server component (DKR) to hold information. Real DKRs require more features than the current very simplified WWW can provide. If a DKR could be built today it could act as a proxy, storing the "meta data" added to content parts that exists out on the Internet. DKRs can grow behind this caching proxy by storing metadata for external content as well as holding content published internally. Another transformation can be developed to serve DKR content out to the rest of the world via the web. The little purple numbers seen on the boostrap.org website are the artifacts of such a transformation process. One extremely advanced model (though still theoretical) for DKR storage is called NODAL. It stands for Network-Oriented Document Abstraction Language.

In essence the OHS can be viewed (my interpretation) as a DKR repository (TWiki) with additional (as needed) features as a web proxy server (like squid). This could allow overlays of greater addressability to existing web content. TWiki uses RCS to allow full version controlled access to content. As the Interface for the versioned content is abstracted (already underway) more powerful versioning engines can be used. Working with Apache authentication is provided for Intranet use. A rudimentary type of this flexibility can be seen in TWiki SkinPackages. The Hyperscope viewing all data is a web browser.

I think of a TWiki system installation as a type of simplified DKR. TWikiShorthand is a text based markup with some similarities to NLS' features. Since TWiki is written in Perl and architected very well, it's designed (like sed and awk before it) to be able to interpret data streams dynamically. Additional features in the form of Plugins and Add-Ons can be installed and upgraded independent of the engine. Some people view the requirements for a real DKR as impossible, yet Doug has pursued this vision to the world's benefit.

-- GrantBow - 15 Jan 2003