ORB Visualization
(soon)
****************************************
It takes a
Community to Create a National Project
****************************************
A descriptive enumeration of the envisioned ORB Peer-to-Peer system is as follows:
GOAL: Provide a remotely operated system for creation/manipulation/viewing and enabling of ORBs and techniques that draw upon the ORB
1: Client-side software is developed
1.1: is open source (PERL)
1.2: is ported to allow it to run on Windows, Linux, *BSD
1.3: includes a web interface
1.4: interface for ORB operations (creation/manipulation/viewing)
1.4.1: knowledge engineering interface used by trained specialists
1.4.2: user interface allows user manipulating of subject matter indicator neighborhoods
1.5: local back-ups and offline operation abilities
1.5.2: import and export processes to allow auxiliary manipulation
1.5.1: import and export processes to allow archival
2: Server-side software is developed
2.1: based on currently developed ORB data encoding modeled after the NdCore software (see notational paper)
2.2: is open source (PERL)
2.3: runs on Linux
2.4: runs as a daemon or in a "transparent server" setup
2.4.1: transparent server means that
other services such as SSH (secure shell) are used as a means for the client to
communicate with the server, instead of opening un-needed ports
2.4.2: possible deployment as part of
GRID network computing
2.4: uses the BerkeleyDB hash table system
2.4.1: The Berkeley DB can be completely replaced with the In-memory Referential Base (I-RIB)
2.4.2: The I-RIB can be replaced, in theory, with a derivation of the Instant Index encoding process
2.5: Specific types of processing of data occurs at on the server-side, in response to web-service type client-side messaging.
2.5.1: A distribution of processing load takes a significant load off of the requirements imposed on the client-side.
2.5.2: Reduces the amount of bandwidth required for complex and large-scale operations.
2.5.3: This is not typical thin-client architecture since the distribution takes into account the needs of the user as well as capitalizes on the efficiency of Berkeley hash tables, I-RIB “keyless” hash tables and the Instant Index encoding innovation.
3: The entire system (server and client: includes
3.1: A Method of Disambiguation over Subject Management Indicator neighborhoods
3.2: A Method of Ambiguation over Subject Management Indicator neighborhoods
3.3: Methods for visualizations and rendering of viewpoints (SLIP)
3.4: Component to export ORBs to SLIP Data warehouse format
3.5: Component to import SLIP results to ORBs
3.6: An ORB subject matter search engine
3.7: NdCore-type temporal visualization of thematic evolution
3.8: Method to provide terminological control and reconciliation processes
4: Allows collaboration on ORB tasks over distributed environments
4.1: may be integrated with the well used Groove collaboration space
4.2: may be integrated within a knowledge management system like SchemaLogic’s SchemaServer
4.3: ORB structured data does not often have first order predicate logics like OWL or Cycorp ontologies, but does provide an inferential process.
5: Integration of other products/system/techniques
5.1: sentence parsing integrated into clients/servers
5.2: Instant Index may be used as a means to create the ORBs
5.2.1: The SLIP is now the completed visualization tool for an ORB
5.2.2: An ORB is the network produced from any set of syntagmatic units, having the form < a, r, b > and a and b are locations and r is a relational operator.
5.2.3: We conjecture that ORB’s can be optimally encoded into the Instant Index encoding and that retrieval of selected elements can be made almost instantaneous – no matter the degree of complexity.
5.3: Output to formats and integration of other system formats (as they arise)
6: Application of knowledge science/management techniques
6.1: taxonomy generation capabilities
6.1.1: provides a next generation full life cycle taxonomy generation capability at low cost
6.1.2: has no J2EE or .NET infrastructure requirement
6.1.2.1: needs only TCP/IP
6.1.2.2: handled correctly, the raw signal is already compressed bit wise
6.2: all current and future requirements draw on the ORB encoding
6.2.1: allows the movement towards neuro and quantum neural architectures
6.2.2: maps well to mature research on behavioral architectures related to selective attention, false sense making and coherence measures
6.3: Can be easily implemented due to the simple nature of the ORB system, it's design, and coding
This seems to be the direction the system is taking right now. Basically, it creates a structure which allows for many people across a distributed environment to work on the same result set derived from text, as well as produce new result sets, and modify/refine existing result sets.
The system then also allows these results to be used within other programs for further study/analysis/refinement, and then brought back in to the collaborative workspace.
Ultimately, programs such as the SLIP program can be quickly and easily updated to enable them to directly interact/import/export with/from/to the ORBs.
A few of the really key elements include the integration of parsing capabilities like those seen in TAI's VisualText IDE, as well as the data collection abilities apparent within the Instant Index (kill two birds with one stone, while indexing, we're also harvesting. This then develops a direct competitor to the ClearForest tool set.
The concept is a collaborative environment that allows for quick and easy expansion as science and business requires.
-- Nathan Einwechter and Paul Prueitt - November 24, 2003