Home What's New
Inmagic Software
Consulting Services
Web Publishing Products
WebPublisher Examples
For IT Professionals
Web Database Hosting
Training Workshops
Bar Code Services
About Crew-Noble
Site Index Contact Us
|
An Overview of Technology: DB/Text |
|||||||||||||||||||||
|
Definition of Textbase Textbase combines the capabilities of database technology (fields, report writer, data entry) with full text technologies (rapid word retrieval, variable and unlimited length data model). Textbase is not an "after the fact" combination of these two technologies, but rather it is designed from the ground up as a database for text. Textbase technology provides an environment for the management of text that is superior to the two primary competing technologies: relational database technology and full text technology. Relational databases are too slow and too rigid to provide a robust environment for text management. Full text products have the speed and flexibility for text but lack the structure, search precision, data management and reporting capabilities of the textbase approach.
What
Makes DB/TextWorks Special?
Most organizations do an effective job of managing numeric information. They know exactly when to order a particular part or raw material ...they know how much to budget for next year's travel expenses, and so on. They have control of this information because they have tools that are designed to manage information that is predictable in size and format, such as dates, telephone numbers, and monetary amounts. Most information is textual But textual information is less "obedient," varying widely from one instance to the next. And text is far more prevalent in the workplace than numeric information. In fact, according to a Gartner Group study, as much as 90 percent of all information is textual. In an effort to manage textual information, organizations have tried either standard database management systems or text retrieval systems. But neither is exactly right for the job. Standard database systems are too slow Because standard database systems do not create keyword indexes, searching for textual information is slow. In fact, a simple keyword search can take literally minutes with a database system. This shortcoming also impacts the type of searches that are possible with a database system. For example, searching for phrases (global warming effects), word stems (warm*), and proximity relationships (global within 10 words of warming) are either too slow or impossible to execute without extensive programming. |
Standard database systems are too rigid Because they have fixed-length fields, standard databases do a poor job of storing textual information. As a result, information must conform to the database, causing users to abbreviate and/or truncate valuable information Although many database products introduced "Memo fields" to try to alleviate this problem, you can still encounter problems. For example, you can't keyword index Memo fields, so searching them is slow. Standard database systems are too "flat" Standard database systems are rigid in another way: they are two-dimensional, structuring information into rows and columns, a framework that cannot deal cleanly with a field that requires multiple entries. For example, suppose you want to track the purchase history of your customers. With a standard database system, you have three options, each presenting a different problem. One option is to "blob" purchases into a single text string, losing all distinction for reporting and sorting purposes. Another option is to guess at the maximum number of purchases a customer will make and create a field for each (e.g., Purchase 1, Purchase 2, Purchase 3). If you guess too low, you'll constantly be dumping and reloading records because you have to add another Purchase field to the data structure. If you guess too high, you'll waste hard disk space because the database allocates space whether it's used or not. The last option, creating relations between tables, complicates database design and maintenance. Text retrieval systems aren't structured enough Looking for more flexibility and speed, organizations have tried full text retrieval systems, but found that the benefits of speed come at the expense of precision and control. Searching is imprecise with full text systems Because text retrieval systems do not break records to the field level, queries cannot be narrowed by searching two or more fields at the same time. For example, suppose you set up a textbase to manage all internal and external correspondence (letters, proposals, contracts) for a large engineering project. You want to find all letters from Bob Jones. Your search in a text retrieval product would find, among other things, all letters to Bob Jones, all memos to Bob Jones, all letters and memos about Bob Jones, etc. A textbase will allow you to specify that you search only the "Author" field, something full text products can't do. Full text systems can't write reports Because fields are the building blocks of reports, and full text systems do not have fields, it's impossible to create even simple reports. With full text systems, you can print what you find, but you cannot perform fundamental report operations such as calculations, sorts, sub sorts, totals, subtotals, and record counts. Nor can you design reports to customize the way information appears. The best of both worlds Like a database system, DB/TextWorks breaks records to the field level, enabling you to write reports, sort and sub sort information, and otherwise manage information with great power and precision. Moreover, users have the added flexibility to add multiple and distinct entries to a field at any time (repeating "3-D" sub fields). You can also make "relational-like" links between two or more textbases to eliminate redundant data. And similar to a text retrieval system, word searches and other types of queries are virtually instantaneous. Because all fields are variable and unlimited in length, you can manage full and abstracted text, as well as other data types, such as dates, images, numbers, computed numbers, and so on. |
||||||||||||||||||||
|
Summary Comparison of Core Capabilities
|
|||||||||||||||||||||
|
Inmagic software | Crew-Noble services
| Training Dates
Crew-Noble Information Services Revised: 10/9/2007 |
|||||||||||||||||||||