Computer Chronicles Revisited 88 — Double Helix, PFS Professional Write, and R:BASE

Database managers often tested the limits of 1980s personal computers. For example, longtime Computer Chronicles contributor George Morrow faced an ongoing problem with the database he maintained to catalog his massive collection of 78 records. He told the final issue of the Morrow Owners’ Review that he’d been forced to abandon his own Morrow Designs MD11 because the old CP/M machine could no longer hold the database.

Morrow initially used a database manager called Personal Pearl, but after about 9,000 records or so, it “got severe hiccups.” So he moved to Ashton-Tate’s dbase II. But after he reached record 32,678 on that program, it “wrapped around on itself and destroyed records.” (This would be the dBASE equivalent of the Pac-Man level 256 kill screen.) Thankfully, Morrow managed to move the database to an MS-DOS machine and repair his data thanks to Norton Utilities, something he discussed in a previous Chronicles episode.

As for this next Chronicles episode from May 1987, the focus was on the successor to dBASE II–dBASE III–and several of its competitors in the MS-DOS and Macintosh database market. The episode’s cold open featured, appropriately enough, Stewart Cheifet visiting George Morrow at his home. Cheifet posed the question, “What do you do when somebody says to you, ‘What do you have by Hoagy Carmichael from the year 1930?’” Either you needed another room full of 4-by-5 index cards and a week of time, or you used a personal computer with a database manager to get the answer in seconds. (Morrow worked in the background and quickly produced the answer to Cheifet’s question, which was “six.”)

Back in the studio, Cheifet showed Gary Kildall several “primitive” forms of database management, including a Rolodex, an address book, filing folders, and an accordion-style file box. But these all had disadvantages. They took up a lot of room, were hard to update, and could only be sorted alphabetically. That’s why many people had now moved to computer database management.

Cheifet noted there were two basic versions of database managers: relational and flat-file. He asked Kildall to explain the difference. Kildall said a flat-file manager basically took a sequence of records, such as names or addresses, and sorted them to make something like a mailing list. That was a very simple operation. In contrast, relational databases were based on a more complete mathematical theory. Every piece of information was considered a table. Each line of the table was called an n-tuple. In mathematics, that was called a relation and hence the name relational system. For example, a Rolodex could be considered a table. You might then have another table listing geographic sales areas. With a relational database, you could extract the names from the Rolodex and put it together with the sales area table to get a new table. You could then keep adding tables to develop a more comprehensive mailing list.

Accounting Firms Turned to Macintosh Database Solution

Wendy Woods presented her first remote segment, which focused on the use of Macintosh-based database management software at the San Francisco office of accounting firm Peat Marwick. Woods said that behind every successful office-bound PC was a software program that either sped up or replaced tasks formerly done by hand. Peat Marwick found a way to do both and create a new communications link at the same time. The firm’s graphic arts department replaced billing sheets with an all-Macintosh network, combining Odesta Corporation’s multi-user Helix database with MacDraw and a LaserWriter printer to order, bill, and create graphic presentations in as little as 24 hours.

The Helix database, running over an AppleTalk network, operated like an electronic assembly line, with the additional feature of moving in any direction or speed required. As work orders were received, Woods explained, they were entered into the database by client account, date of entry, and subject description. The client’s rough sketch was digitally reproduced and stored on disk for later printing. The file moved through different stages, accumulating charges automatically according to the time spent on each task. The network permitted each authorized user to verify the completion of each step before the final printout.

Looking at Flat-File Databases

Dennis Travins and Jim Hubbard joined Cheifet and Kildall in the studio. Travins was a video production manager with the Stanford Instructional Television Network (SITN), where he worked with Software Publishing Corporation’s PFS Professional File. Hubbard was a Real Estate broker who used Symantec’s Q&A.

Kildall noted their guests were users of database systems rather than software developers or marketers. How exactly did they use database management in their particular line of work? Travins said that at SITN, they broadcast live computer science and engineering courses out to Silicon Valley directly from studio classrooms on campus. His job was to manage six different studios throughout the day, keeping track of all of the production staff and professors. He used a database to accomplish this and manage the necessary accounting.

For his part, Hubbard said he used a database to keep track of the property owners that he worked with. He had an area of 500 to 600 people and entered their information–names, addresses, telephone numbers, et al.–into the database. He also used the database to do mailings.

Cheifet asked Travins to demonstrate PFS Professional File and explain why he used that program. On an in-studio computer, Travins showed what he called an “outline of a database” that he used to keep track of all of his people and studios. For instance, the database tracked the time-sheets for each student who worked at SITN. A student could enter their information offline. Travins then totaled them up using the database. Another feature of PFS allowed him to create reports. Using a pre-developed macro, he pulled up a cross-tab report that showed the cost of putting on a specific production at SITN.

Kildall asked Travins how many records he kept in the database at one time. Travins said he’d had as many as 365 active records at once. Stanford ran on a quarter system, so there were about 300 to 400 records per quarter. He added that one reason he liked PFS Professional File was its integration with other PFS software. This enabled him to dump information from Professional File into the matching PFS spreadsheet.

Turning to Hubbard, Cheifet asked him why he used Symantec’s Q&A in his real estate brokerage. Hubbard said Q&A was a very powerful database manager. He kept information on between 2,000 and 3,000 properties in his database. This included data about each property–lot size, number of bedrooms, et al.–as well as contact information for the owners. Q&A allowed Hubbard to access the database, gather information, and do a mailing.

Kildall noted that Symantec advertised Q&A as having an “artificial intelligence” component–i.e., it was able to parse natural language. Kildall asked Hubbard for a demonstration of that feature. Hubbard pulled up some forms he used for a prior mailing in Q&A. He entered: SHOW US THE FORMS MAILED ON 4/2/87. The program prompted him to explain what he meant by FORMS. Was that word part of the built-in dictionary or was there some other information required? The software also didn’t understand the word MAILED. Hubbard said there was likely confusion because the field used in the actual database was DATE MAILED. So if he changed his inquiry to say DATE MAILED it would probably work. (Basically, the “natural language” sentence had to use words defined in the database.)

So Hubbard entered a new inquiry: SHOW US THE FORMS DATE MAILED ON 4/2/87. Once again, the program got stuck on the meaning of FORMS. Cheifet quipped that the software didn’t actually learn anything from the previous query. But after Hubbard re-explained the term, Q&A managed to produce a report showing the street number and street name for all forms where the date mailed was April 2, 1987.

Notwithstanding this particular demo, Cheifet asked Hubbard if he was pleased with how Q&A worked in his real estate practice. Hubbard said he was. He noted that Q&A came with a nice word processor. And of course, the “artificial intelligence” was nice to run particular reports. He ran another demonstration–RUN THE MAILING REPORT–which went off without a hitch. This produced a report that showed all of the properties that he’d mailed to with the date of each mailing.

Optometrist Eyed Own Database Solution

Wendy Woods returned for her second and final remote segment, this time featuring a San Bruno, California, optometrist, Dr. Gordon L. Epstein, who developed his own custom database manager to run his practice. Woods explained that Epstein had tried, but wasn’t satisfied with, any of the software designed specifically for optometry. So he put together his own, called Eye-Base.

This was actually a customized interface built on top of Ashton-Tate’s dBASE, Woods said. Epstein used Eye-Base to track inventory, patient accounts, and appointment scheduling, among other things. Epstein told Woods that the computer was a means in which he could make the rules. And if they don’t obey those rules, he could make changes. Indeed, Epstein had added other programs to his custom system, including Paperback Software’s VP-Info. This was a graph program that allowed him to compare the time it took to receive supplies from different vendors and determine which activities generated the most revenue.

Woods said that as with many businesses, optometrists were at a crossroads. Caught between changing technology and the limits of various software packages, there was a great deal of confusion. That was why Epstein and thousands of others had to become as good in computers as in their own professions.

Looking at Relational Databases

Thom Kozik and Lynn Luukinen joined Cheifet and Kildall for the final studio segment. Kozik was the applications director with Ashton-Tate, the developer of dBASE III Plus. Luukinen was director of advanced products with Microrim, which developed and published R:BASE.

Kildall opened by noting the products demonstrated in the previous studio segment, PFS Professional File and Q&A, were both flat-file database managers. How did a relational database like dBASE differ? Kozik said a relational system took advantage of the fact that a lot of data was often kept in different departments or disparate sets of information. So you could relate information from one file to another, such as connecting invoices to orders. There was a common number generally kept between both pieces of information that made them related.

Cheifet asked Kozik for a demonstration of dBASE III Plus. Kozik showed off the menu-driven assistant that came with the new version. This assistant let users program settings for commonly used files and tasks into a “view file” type that could then be accessed directly from the menu. The user could thus restore their settings with a single command.

Continuing the demo, Kozik said if you knew the piece of information you wanted to look for–say a specific customer number–you could enter it directly into a search field. The software then positioned you at the appropriate record. A status line at the bottom of the screen informed you of your current location within the database. (dBASE didn’t actually display the record at this point, it only informed you that it had located the record.) Kozik then selected the EDIT function to look at the record inside a pre-created screen. You could use the assistant to create additional screens for the same database.

Kozik added that dBASE was designed to handle millions of records at a time. It was obviously hard to work with that many pieces of information. So you could filter down information by creating a query. Using an “advanced query system,” you could create a custom search. For this demo, Kozik created a query to look for records that used the word “group.” This allowed you to work with a smaller set of data. dBASE also kept a record of previous commands entered in a history buffer, which could be saved to a different file.

Cheifet turned to Luukinen and asked about her company’s product, R:BASE, which was also a relational database manager. How did it differ from dBASE? Luukinen said the primary difference was that R:BASE provided additional tools with the database to allow the users to automate functions. For example, there was a tool to automate applications. There was also a tool to create complex forms or reports without the need for programming.

Cheifet next asked about the improvements in the most recent release of R:BASE, known as System V, over the original product. Luukinen said there had been lots of different changes. The aforementioned automation tools were probably the biggest change. There was also now multi-user capability and new built-in financial functions.

Luukinen then provided a demo of R:BASE. Specifically, she demonstrated the application creation feature. She showed how you could modify a previously created application designed to create reports. Essentially, this allowed you to create and write an entire macro program without having to learn the underlying programming language or write code. The new application was then available from the main menu screen.

Kildall pointed out that IBM recently announced a long-term plan to create its own relational database system based on the SQL standard. What effect would that have on the guests’ companies? Kozik said that Ashton-Tate planned to continue down its own path of developing for the 8088- and 8086-based machines while also working with the newer 80286 and 80386 platforms. Luukinen said that IBM planned to integrate its database into the operating system, and that the tools developed to create databases within R:BASE could possibly be adopted for IBM’s structure as well.

IBM Slashed XT Prices to Clone Levels as New PS/2 Hit Shelves

Stewart Cheifet presented this week’s “Random Access,” which was recorded in May 1987.

Niche Databases Endure Into 2023

In the previous posts on this business applications miniseries, I discussed how the markets for word processors and spreadsheets followed a similar pattern. An early market leader from the pre-IBM PC days was replaced by another dominant application in the MS-DOS period that was itself replaced by Microsoft once Windows became widely accepted. Database management software, however, proved a bit different.

While Microsoft Access certainly gained prominence once we get into the 1990s, database software as a whole retained a lot more of the smaller niche applications. Indeed, two of the programs featured in this episode, R:BASE and Helix, are still in active development today, 36 years after the original airdate. Two others, dBASE III and Q&A are long gone, as I’ve discussed in earlier posts.

A Brief Overview of Software Publishing Corporation

A third, PFS Professional File, is also no longer with us. The developer, Software Publishing Corporation (SPC), has been mentioned several times on Chronicles up to this point, with Paul Schindler reviewing a few of their products. Fred Gibbons, Janelle Bedke, and Job Page founded SPC in 1980. The trio previously worked at Hewlett-Packard. Page recalled that one day in 1979, he ran into Gibbons in the hallway at HP. Gibbons worked in marketing, while Page worked on database managers for minicomputers.

Gibbons brought up VisiCalc, the spreadsheet program recently released on the Apple II. It led him to believe there was now a potential market for standalone microcomputer software. He gave Page an Apple II to play around with. Page eventually developed a database manager called Personal Filing System, later abbreviated pfs, which became the brand name for a host of SPC releases in the business applications market. This included Professional File, which was meant to compete with Ashton-Tate’s dBASE, then the market leader.

As I mentioned in a recent post, SPC acquired Wisconsin-based Office Solutions, the publisher of the word processor OfficeWriter, in 1988. Office Solutions then effectively became SPC’s customer support center. SPC itself undertook a major downsizing in 1991. Gibbons decided to sell the entire pfs line to Spinnaker Software. This was part of a shift away from consumer software to products focused on the business market, notably the graphics presentation program Harvard Graphics, which SPC first released in 1986.

Unfortunately, sales of the DOS-based Harvard Graphics started to plummet as Windows 95 approached. Gibbons resigned as CEO in 1994. Two years later, in October 1996, Allegro New Media acquired SPC in an all-stock transaction. I’m not exactly sure of the fate of Professional Writer. It likely went to Spinnaker as part of the 1991 sale, and Spinnaker in turn was absorbed into The Learning Company in 1994.

R:BASE Quietly Chugging Along Under Longtime Developer

We also have the usual bit of corporate ownership shenanigans with our two surviving products, R:BASE and Helix. Wayne J. Erickson founded Microrim, the original developer of R:BASE, in 1981, together with his brother, Ronald Erickson, and investment banker Andrew Evans. Wayne Erickson previously developed relational databases for Boeing Computer Services and the University of Washington.

As was the style at the time–and now, for that matter–Erickson made an early push to market his database products as featuring “artificial intelligence.” He told Knight-Ridder’s Mary A.C. Fallon in 1984, “The AI market is really just starting; it might be 2001 before we get there,” adding that AI would eventually replace the mouse and desktop icons as the common user interface for computers.

While then–as now–this AI talk was mostly industry hype, Microrim proved successful with its R:BASE product. In June 1986 the company released R:BASE System V, the version demonstrated in this episode. The most recent version, R:BASE X.5 (or 10.5) came out in October 2018.

At some point in the 1990s–I couldn’t nail down an exact date–Canada-based Abacus Software, Inc., acquired Microrim. In March 1998, Abacus announced plans to shut down Microrim as part of a consolidation. A few weeks later, in June 1998, a newly formed Pennsylvania company called R:BASE Technologies, Inc., acquired ownership of R:BASE from Abacus. A. Razzak Memon, a longtime R:BASE developer, was the principal behind R:Base Technologies. Memon, a surgeon by training, started working with R:BASE in the 1980s when he needed a database to assist in his AIDS research. Memon continues to run R:BASE Technologies as of early 2023.

Big Giant Donut Bringing Helix Into Apple Silicon Era

Helix, the Macintosh database manager featured by Wendy Woods in this episode, is also still alive and kicking, making it one of the longest continuously developed Mac apps of all time. Indeed, Helix can trace its lineage back to the earliest days of the Macintosh, first coming out in 1984. Double Helix, the version featured on Chronicles, was a 1986 release with additional features.

Illinois-based Odesta Corporation was the original Helix developer. In 1992, Texas-based Albara Corporation acquired Helix from Odesta and continued to develop the database through its Northcon Technologies subsidiary, which adopted the name Helix Technologies. In 1998, a California-based company, The Chip Merchant (TCM), acquired Helix Technologies from Albara.

TCM filed for bankruptcy in April 2003. Several months earlier, TCM had turned over day-to-day operations of Helix Technologies to Matthew Strange and Gil Numeroff. They would eventually join with a third investor to create QSA Toolworks, LLC, which became the new developer of Helix. In September 2022, a New Mexico-based company, Big Giant Donut, Inc. (BGD), acquired Helix from QSA Toolworks. Dana Bernard, the principal behind BGD, was like R:BASE Technologies’ A. Razzak Memon a longtime user of the database who decided to just buy the company outright. As of February 2023, BGD was in the process of re-writing the 32-bit Helix code into a 64-bit application capable of running on newer Apple Silicon Macs.

Notes from the Random Access File