Classic Computer Magazine Archive START VOL. 1 NO. 4 / SPRING 1987

ST Database
Management Systems

by Brian Lee

The ST has been in developers' hands for two years now, and is finally being blessed with database programs. Just how good are they? START's business expert Brian Lee takes a look at ten ST programs designed to help you keep track of what goes where.

The term database management system, or DBMS, has been widely used since the advent of computers. DBMSs have long been used to organize and maintain large collections of business information. However, most of the early DBMSs were designed specifically for large mainframe computers. Smaller businesses could not afford to use the power of these DBMSs until the introduction of powerful personal computers and products like dBASE.

There are now a variety of databases available for the ST-some good, some bad, with prices ranging from $50.00 to $150.00. In this article, I will present a quick summary of ST database features, which I hope will shorten your head-scratching time when you next visit your local computer store and look for an ST data management program. At the very least, this article should enable you to ask the right questions before you make your purchase. But first, some background material is in order.


The term database simply refers to a collection of useful information organized to facilitate easy retrieval and processing. Personal computers such as the RAM-abundant ST, with the horsepower of the Motorola 68000 chip, offer a cost-effective way to store relatively large amounts of information while providing flexible and fast access to data.

Databases can range in complexity from a simple name and address file to complete accounting and inventory control systems. A simple name and address file could include the following data items:






The data stored in this data base could be organized as follows as illustrated in Figure 1.

Each piece of information, for example first name, is called a field. Each row is a collection of fields of data and represents the information for one person. This collection of fields is referred to as a record. Simply put, a database is a collection of records.

The above list, however, may not be the most systematic approach to organizing your records. For instance, we would probably prefer to view the records in alphabetical order by last name or perhaps sorted by zip code. This rearranging of data is typical of the flexibility offered by the DBMS.
Figure 1: Sample mailing list
Gordon Wong 434 Eyglass rd. Mill Vally, CA 94054
Mike Silva 787 Forte Ln. Oakland, CA 94610
Roy Wolford 123 Paco Ave. Corte Madera, CA 94930
Allen Orcutt 943 Ronco Dr. Vegamatic City, CA 95065


All information contained in a database can be called data. Data can be categorized by its nature and use in applications. Generally, data can be classified into two broad groups: alphanumeric and numeric. Some DBMSs use additional categories and subcategories for special data like dates or pictures.

Alphanumeric data consists of alphabetic characters (A-Z), numerals (0-9), and common symbols like punctuation marks. Alphanumeric data is treated as text by the DBMS and cannot be used for computations.

Numeric data represents a quantity. It is treated by the DBMS as a value which can be used for computations and included in mathematical formulas.


The information stored in a database can be organized in many ways. The most conventional models of organization are hierarchical and relational models.

Hierarchical data bases are rigidly structured, organizing their contents in a configuration resembling a tree. (A good example would he a flowchart.) This tree not only defines the data elements but also the relationship between the elements. While these tree structures can vary in complexity. they are always characterized by their well-defined structure. Contents of the data base are inextricably linked to the structure. A common example of a tree structure is the typical organizational chart shown in Figure 2.

should enable
you to ask the right
questions before you
make your purchase.

Most hierarchical databases contain their information in a single data file. File managers reflect this characteristic, indicating their reliance on a single disk file as the major structural element. Within this file, data elements can be grouped by common characteristics (as in the Marketing Department, in the above example). While hierarchical databases are conceptually easy to deal with, they are severely limited by their ability to change their organizational structure depending on the needs of the user.

File managers are generally simple-to-use, highly effective tools for managing personal databases like addresses, recipes, tape collections, and the like. They are well suited to virtually all home database applications.

On the other hand, relational databases organize data into collections of two-dimensional tables consisting of rows and columns. Each row contains information for one record in the table. Within each row, information is divided into separate fields. As a result, each data element can be referenced by a unique row and column location within the table.

A single database can contain many of these tables with each table existing as a separate data file. Data in different tables can be linked by establishing a relationship between the tables based on matching data. This is the key feature of the relational model; the separation of the data itself from the structure of the database. The relational model permits the easy restructuring of the database by simply redefining the network of relationships between tables without affecting the contents of the tables. Truly relational data bases are complex programs and, as a result, challenging to make user-friendly. In the IBM market several products have made major steps towards simplifying the use of relational data bases. In particular, R:BASE System V has created a completely menu-driven shell which allows the user to visually define and edit data tables.

Figure 2

A common application of relational data bases is accounting systems. A business will generally have a collection of regular customers, each with their own address, credit terms, etc. When the business receives an order, an invoice must be prepared. To do this, information about each of the items being sold as well as the customer information must be gathered. A file manager would require all information for each invoice including customer address and the information for each element purchased to be included in each record. As a result, valuable disk space is wasted carrying redundant data. The relational model overcomes this by allowing customer addresses to be contained in a separate table referenced by the customer's name. In addition, the information for each of the items being sold (description, price, etc.) can be maintained in a separate table referenced by a part number. Under this model, the invoice only needs to store the customer name and part numbers to complete the invoice.


For purposes of simplicity here, I will use two categories for my brief database reviews: File managers and Relational DBMS. (I'll generalize the term file managers to encompass all non-relational products.) While some may dispute my definition of relational database managers, I will use the most rigorous meaning. To be considered a relational DBMS, the product must maintain separation of data contents from the structure of the database to facilitate flexible and dynamic restructuring of the database.

Data can
be organized by
its nature and use
in applications.

The most important criteria for deciding which database product is right for you is to define your objective carefully For most applications a file manager will give you the benefits of ease of use and simple management of your data. If you must manage large amounts of complex data and are willing to invest additional time and money, the relational products reviewed here will give you the power necessary.

I have reviewed several products in each category. All the databases share common features, but I will focus on those features which make the product unique. The START charts at the end of the article will provide more detailed information for product-to-product comparison.

MailList from Artworx

MailList, as its name implies, is designed to manage a name and address file. The program supports one key field and five data fields of up to 32 characters. The program will handle up to 1800 records on a single-sided disk. MAILLIST's user interface is by command line, making no use whatever of GEM or the mouse. The documentation consists of one small double-sided sheet of paper containing a "tutorial." While this tutorial takes you through some of the functions of the program, there is no explanatory text accompanying the directions for what keys to press. Even considering its low cost, this program is not up to commercial software publishing standards.

Datatrieve from Abacus

Datatrieve is a well-designed product which makes good use of the GEM interface and also gives you quick access to all functions from your keyboard. The program allows you design screen and report masks which can include graphic elements like boxes and shaded areas as well as font specifications. In fact, you can even specify different fonts for the field names and field contents. Datatrieve allows records of up to 64000 characters and text fields of up to 32000 characters. Files can be indexed on up to 20 fields and searched using wildcards, substring matches, and ranges. Reporting is somewhat limited to labels and lists with a single-level break. Page headers and subtotals are possible. The 123-page documentation is quite good, with numerous illustrations.

dbMaster One from Stoneware

dbMaster One consists of three separate programs: one for creating a database, one for maintaining a database, and one for data import and export. The program makes good use of GEM for the creation of the database, allowing the user to stretch fields to the desired size. One of the unique features of the program is the support of multiple lines of data for a given field. The lack of integration between the three program modules is a bit of a nuisance but not a major drawback.

HabaView from Haba Systems

HabaView uses a columnar format to display data. This model is very much like a spreadsheet, with each row corresponding to a record and each column representing a field. The program also makes very good use of the GEM interface. It is easy to stretch a column to a new length by dragging its right edge, and columns can be reordered easily. HabaView also provides the option of viewing the data in the more conventional "form" view. Fields in the form can be easily resized and moved. The program also has some nice touches like optional automatic capitalization of the first letter in the data, which can save time when entering names and the like. Multilevel sorts are supported through use of "progressive" sorts. Lists and labels are supported but the user cannot create flexible custom report formats. The 58-page documentation is quite good at explaining the operation of the program. Overall, an easy-to-use program but one which might limit the more ambitious user.

Zoomracks II from Quickview Systems

Zoomracks II is a radical departure from conventional database programs and does not fit neatly into a category. Zoomracks II is based on the analogy of cards in racks, like those found next to timeclocks. The program does a good job of insulating the user from the usual concerns of field types, lengths, and file handling. The documentation and accompanying reference card and keyboard template are also well designed and easy to follow. Several pre-designed applications are available as separate products. One of the most unique features is the program's ability to include DEGAS picture files in the data base. This is clearly a product driven by a singular vision, which provides an innovative approach to data management. The product provides strong functionality at relatively little effort on the part of the user. First-time-users of databases are likely to find the rack analogy more intuitive than those of us prejudiced by experience with more conventional products.

Data Manager ST from Timeworks

Designed to integrate with Swiftcalc ST and Word Writer ST, Data Manager ST is a powerful and well-designed product. It makes full use of the GEM interface as well as the function keys, a reference for which is displayed on screen. The program uses a columnar list format like that of HabaView and also provides a form view. However, Data Manager ST goes a couple of steps further by providing calculated columns and flexible custom reports. The user can re-size and relocate columns, although moving a column requires more keystrokes than with HabaView.

The most
conventional models
of organization are
hierarchical and
relational models.

Data Manager ST provides flexible report creation and even business graphics in the form of bar graphs, line graphs, and pie charts. The 161-page manual is extremely well done and easy to read. This is an excellent and professionally done product which does an admirable job of balancing power and ease of use.

The Manager from BMB Compuscience

The Manager is a menu driven file manager It uses standard item list type menus which you can select by number. The program does not make use of GEM, nor does it use the mouse. Yet the key to this product is its powerful programming language, Manager Math. Manager Math is actually a procedural language which can handle data manipulation and even multiple file operations. Functions included permit not only calculations but also string manipulation. This permits the creation of custom reports and data entry screens. The excellent documentation which accompanies the program is very thorough and replete with examples. If you are familiar with file managers on other computers like Apples or IBMs, you should feel right at home with this product. This is a polished product which looks like a conversion from the IBM (it even uses the numeric keypad for PgUp and PgDn just like IBM!).

Trimbase from Talent Computer Systems

Trimbase utilizes GEM and the familiar concept of a card file to represent data. The program purports to be a fully relational database management system, providing functions for merging and joining disparate data files into new files. However, this one-time processing of files accomplishes only a snapshot of the relations between the data files. Updating of information in the source files is not represented in the destination file until the joining function is completed. Trimbase includes a macro recording facility to expedite this process. Overall the product provides fairly easy access for the non-programmer to relational functions, but would be less appropriate for development of custom turnkey systems.

dBMan from Versasoft

dBMAN is a serious database management system. Compatible with the industry standard dBASE III, dBMAN provides the custom database system developer with the flexibility and power necessary to develop turnkey systems. The program supports command files (though limited to 236 characters), local and global memory variables, and permits flexible report creation. The documentation is a well organized reference to the commands and functions of the program. The product is clearly aimed at experienced database designers. There is no provision for menu-assisted applications development or report generation like those found in dBASE III Plus or R:BASE on the IBM. Also, the user interface is by command line only. There is no link to GEM by applications developed using dBMan. However, the company is reportedly working on a GEM interface at the time of this writing.

Regent Base from Regent Software

Regent Base provides both a GEM-based interface and procedural language for applications development. The program provides a unique form design capability which includes action buttons. You can make these buttons perform predetermined functions by activating them with the mouse. Several preset applications are included, ranging from mail lists to check registers. The documentation is sketchy and void of illustrations, a problem which Regent is addressing in version 1.1. Regent treats data internally as tables with a full complement of relational operators. Regent Base, like dBMAN. is suitable for custom database applications development. It is clearly targeted at the ST market. providing access to GEM for programming application screens. lts syntax is somewhat unconventional, a problem which will be mitigated by the revised manual.


The Atari ST market is still relatively young. At this point, many of the database products available lack the power and polish found in similar products in the IBM market. To put this in perspective, however, you must remember the IBM PC software market has been in existence since 1981, and can thus be called "mature." As the ST establishes its own standards of (dare I say it?) "Power Without The Price," we can expect higher-quality, more greatly polished software to hit the market, from both the above manufacturers and others. While the IBM market continues to set the standards for databases, the ST database products have the potential for creating completely new standards- their own.


  • MailList
    150 N. Main St.
    Fairport NY 14450
    (716) 385-6120

  • Datatrieve
    Abacus Software
    P.O. Box 7219
    Grand Rapids, MI 49510
    (616) 241-5510

  • dbMaster One
    Stoneware (distributed by Atari Corp.)
    Atari Corp.
    1196 Borregas Ave.
    Sunnyvale, CA 94086
    (408) 745-2000

  • HabaView
    Haba Systems
    6711 Valjean Ave.
    Van Nuys, CA 91406
    (818) 989-5822

  • Zoomracks II
    Quickview Systems
    146 Main Street, Suite 404
    Los Altos, CA 94022
    (415) 965-0327

  • Data Manager ST
    Timeworks, Inc.
    444 Lake Cook Rd.
    Deerfield, IL 60015
    (312) 948-9200

  • The Manager
    BMB Compuscience Canada Ltd.
    500 Steeles Ave.
    Milton, Ontario L9T 3P7
    (416) 826-2516

  • Trimbase
    576 South Telegraph
    Pontiac, MI 48053
    (313) 334-5700

  • dBMan
    Versasoft (distributed by Atari Corp.)
    Atari Corp.
    1196 Borregas Ave.
    Sunnyvale, CA 94086
    (408) 745-2000

  • Regent Base
    Regent Software
    7131 Owensworrh, Suite 45A
    Canoga Park, CA 91303
    (818) 882-2800