How Verity's TOPIC , when combined with Data Warehousing, Can Manage Information


O Business systems objectives

"Computing is Competing". That must be the focus for the use of computers in the business organisation. At both HQ and in branches and affiliates, computing resources need to be used effectively to improve business competitiveness. The objective is a computer system which supports the organisation's work process. Data is entered only once by its originator with information available to those who need it. Computer systems carry out work for users and communicate with each other and users.

O Business Systems Needs and Computing Approach

In most business, computing needs to reflect the operational nature of the business. Business requires quick actions to respond to market activity and has difficulty forecasting the future. This requires systems to be developed quickly with not very clearly defined knowledge of specifications, and after first being deployed as prototypes, cthen hanged frequently as users better understand their use. The use of scenarios to explore the insights of history are also a key. And this needs common historical data sources.

Thus the systems group should practice business ownership and control of systems, with the business driving technology. Business management of common data and systems promotes effective integration with business operations. Business users should be encouraged to own, develop, and use systems the maximum extent possible. The systems objectives should be that all users in all organisations (HQ, branches and affiliates) have transparent access to common information and systems. Decision makers (regardless of level) are given prompt access to distributed information and systems, and the end-users are provided with effective access to centralized information and systems.

O Business Systems Problem

Business users (both in HQ and branches) work now in a collaborative manner, but this involves much repetitive data entry into separate computer systems, and person-to-person electronic mail, fax and telephone calls. Significant time is spent in manual activities that computer systems could do, thus freeing staff time for thinking and planning. A system that supports this collaborative activity in an efficient manner would be a workflow computing system.

A workflow computing system takes advantage of four trends in computing:

- integration of messaging in all applications,

- applications that improve the flow of work through an organization, and

- use of 'agents' for messaging, and,

- remote access or 'intermittently connected' networking.

O What is workflow?

Workflow computing is AUTOMATING THE TRANSFER OF INFORMATION TO SUPPORT THE FLOW OF WORK. The goal is to create a "just-in-time information" computer system which parallels the organisation's work process. This contrasts with the automation of single, isolated tasks.

The key to implementing workflow computing is the integration of operating text and data with mail enabling. Mail enabling means that business events should tell users that they have occurred and that the action to be taken next should be facilitated by avoiding repeat data entry.

Workflow computing systems must be scalable (that is, easily made smaller or larger) to meet a range of organisation sizes, inclusive so that all who are in the information flow are included, and, be very flexible. The system components are:

- a reliable organisation-wide electronic mail system with the ability to send and receive electronic mail and files to and from third parties,

- LANs in both HQ and all key branch functions,

- a LAN database and line systems group that can build database systems that used to be the province of the mainframe, and

- state-of-the art LAN-based real-time text management system that covers handling, automated receipt, distribution, and filing, as well as ad hoc retrieval.

O Data Warehousing

Uses data management tools to manage operating data. Functions that should be provided are:

- Online summaries of current operating data with simple analysis capability

- Concurrent input and retrieval of operating data

- Programmed and ad hoc reporting on operating data to support business functions

The tools used should be LAN based (internal, intranet, WAN or internet) as appropriate. The LAN systems objective is a computer system which supports the organisation's work process:

- information is entered only once by originator,

- information is available to all who need access, and

- computer systems carry out work for users and communicate with each other and users.

LAN data warehouse systems should be based on the organisation needs, market and workstyle. Choices are OLAP, Relational, SQL or Object oriented. Hardware platform or language/OS should not be a deciding factor and mixed systems can be managed if the appropriate systems management infrastructure is put in place.


O TOPIC Operating Text Management

"Operating" text is text or mail that is related to business data and thus capable of being routed based on business functions or needs and then tied to business data management "keys". It may also contain data for the "data warehouse". It is not personal text or mail. News related operating text is mail that is addressed to the organisation (users can be at all levels), Mail related operating text is mail addressed to specific groups or individuals in the organisation (and can be read by others to meet their business needs).

This operating related text is delivered to a user using TOPIC, in a form prearranged to meet that users needs. It is delivered electronically (with view and print capability) and is organized based on user supplied categories. This operating related text is then arranged, within each category, in order of relevance to the users. The operating related text is also integratable with electronic mail and data systems.

- This new form of delivery is uniquely suited to electronic mail related text.

- It provides a solution to "information overload" since it permits user customization and thus minimizes user absorption time

- It integrates with other information systems, especially those that deliver "workflow computing".

Mail related text System Requirements

- Provide continuously, i.e., in "real-time", categorized mail related text arranged into user specified categories. Categories are those relevant to user's business or personal needs, e.g., geographical, political, business function, environment, financial, etc.

- The mail related text categorization results are ranked in order of relevance with less relevant mail related text not being presented

- Categorization is defined by the user and is based on relevant words and phrases. These words and phrases are associated in loose groupings with optional weights defined by the users. The weights indicate degree of relative relevance. The words and phrases can be imported from the mail related text itself or other systems.

- The word and phrase categorization groups are subject to easy modification

- There is no limit on "richness" of words and phrases

- Multiple mail related text feeds can be handled

- Results can be integrated into other systems, such as "trading" or general data systems and electronic mail systems

Difficulties with "keywords and Boolean" based systems

- No ability to rank search results and eliminate less relevant items

- Not easy to handle rich and complex categorizations necessary for "real-life" use

- No ability to specific relative importance of individual words and phrases

- "Boolean" expressions difficult to handle when complex relations used i.e., many AND, OR etc.

- Requires administrative maintenance of keywords and appropriate filing

How Verity's TOPIC makes this process easy

Consider this simple example of a TOPIC based static search. You have a box of 1000 papers. Each is 1 page with a story on animals or birds or fish. You want only those stories on cats.

You put together list of words and phrases that indicate cats. Examples:

- High relevance - Purrs, Whiskers, Burmese, Tabby

- Medium relevance - Domesticated, Four legs, Tail, Fur

- Low relevance - Animal, Claws

Verity's software allows this TOPIC (or word and phrase collection) to be used to virtually instantaneously retrieve those stories about cats which are then presented in a relevance ranked listing.

Most important for a mail related text systems, this categorization can be done in "real-time" as new papers are added to the box, so that the results list is always being "refreshed".


O How can "mail related text" systems can be implemented

This can best be done by giving the user a mail related text interface that uses TOPIC to do these categorizations in "real-time". This interface should have the following capabilities:

- Show mail related text groupings based on user TOPICS. These are based the above word collections with their optional weights

- Each mail related text grouping provides continuously refreshing lists of mail related text stories in chronological order that meet a certain relevance criteria to the mail related text category

- Provide mail related text story viewer that allows - viewing, printing and forwarding (via e-mail). This can be done on individual stories or in batches. The latter can be used to produce an electronic "clipping" service.

- Multiple mail related text sources can be handled as well as interfaces to other mail or data systems.

This interface is easily built in standard GUI software with the necessary Verity systems interfaces.

In addition, Verity's software can be used for ad hoc searches of historic mail related text material from the document mail related text file.

Advantages:

- Results almost always very relevant

- Handles "information overload"

- Integrates mail related text with other business systems

- Systems requires less manual management, - E.g. keyword management and filing

- Search criteria easy to specify and change

- Searching done in real time

Most important, these TOPICS or word and phrase groupings become building blocks for other TOPICS. In effect, they are portions of the user's knowledge about their needs. Thus these TOPICS are reusable and become valuable in and of themselves.

Disadvantages:

- TOPICS harder to explain than "keywords"

- System requires more systems expertise to manage than Boolean based systems


O Information Management produces Workflow Computing

Workflow computing must involve the integration of a database approach with electronic mail. Integration of operating text ( mail related text that can be tied to business operations ) and data with mail enabling is the key to an effective workflow application. Mail enabling means that business events should tell people that they have occurred, and, where possible, facilitate the action to be taken next by avoiding repeat data entry. In addition, periodic access to information (both read and update) by branches and other non HQ located staff will become part of the daily work process. In a very real sense they are "intermittently connected" to the system.

The information management strategy should cover both data and operating text

Data is managed by the LAN database - OLAP, relational, SQL or "object oriented".

Operating text acquisition:

- delivered by external mail related text feeds or mail system,

- sent to outside business functions by mail system.

Acquired operating text management - routing and retrieval uses "real-time" text search/retrieval engine (based on user managed criteria - words and phrases that indicate business relevance).

Two types of operating text are managed:

- General use operating text, e.g., "news related text" which are automatically categorized into "business relevant" categories for any user use, and

- Business function use operating text, e.g.,., "Mail" which are automatically routed to specific recipients (both prime and subsidiary) by business function.

A system to manage general use operating text can be termed a "personal mail related text" system. "Personal mail related text" is regular mail related text, delivered to users, in a form prearranged to meet their needs. It is delivered electronically in "real-time" (with view and print capability) and is organized based on user supplied categories. This mail related text is then arranged, within each category, in order of relevance to the users, with less relevant mail related text not being presented. In this system multiple mail related text feeds can be handled, and results can be integrated into other systems, such as "trading" or general data systems and electronic mail systems. In addition, the system can support ad hoc searches of historic mail related text material from the document mail related text file.

In the mail system business users can acquire and handle (read, print, annotate) business function operating text "mail" in business function "queues". The "queues" are based on the business functions carried out by users and mirror those words and phrases that a clerk would search for in order to route the mail manually. These queues act as electronic inboxes, providing routed copies of the user's mail for their processing. Reassignment of the few items of mis-routed "mail" is easily done in a cooperative fashion by the users or a dedicated clerk. "Urgent" mail can be flagged for special attention, even automated notification after hours if the user has remote connection to the system. These queues can be reassigned based on business function changes, either manually or using a program and a clerk. Mail "Rerouting" with "post-it notes" is also automated, merely by reentering annotated text into the system, where it is again rerouted automatically by business function. In this system prime users directly manage their operating text, and subsidiary users have direct access to prime users' text.


O Integration of data and text - the Workflow Objective

This provides automated text filing with automatically added data keys. This requires pre-analysis of text and access to data in the data warehouse. When implemented it provides the key ability to go from data elements to relevant text or vice versa. The value of this data and operating text interconnectivity in implementing workflow cannot be underestimated. Consider the effort involved in doing this now. When this is fully implemented, all information is "mail enabled". We then have a workflow system which supports the organisation's work process. This means that the computer system is focused on helping the organization compete in its chosen markets.