Ashton-Tate (Ashton-Tate Corporation) was a US based software company best known for developing the popular dBASE database application. Ashton-Tate grew from a small garage-based company to become a multinational corporation with software development centers spread across the United States and Europe. Once one of the "Big Three" software companies, which included Microsoft and Lotus, the company stumbled and was later sold to Borland in September 1991.
Sometime that year a customer requested that SPI sell a new database program called Vulcan. This led Tate and Lashlee to seek out the program's author. Wayne Ratliff had written Vulcan to help run an office football pool at his day job at JPL. Vulcan was a database application that ran on the CP/M operating system, and was modeled on JPL/DIS, a Univac 1106 program used at JPL and written by fellow programmer Jeb Long. Tate and Lashlee offered Ratliff royalty payments in exchange for exclusive marketing rights for Vulcan, and Ratliff agreed. As a result of this agreement, SPI also became a publishing company.
SPI turned to Hal Pawluk for assistance in marketing and merchandising the new product. Pawluck created the name for the new publishing company by combining George's last name with the fictional Ashton. (Later, George Tate bought a parrot and named it Ashton.) Pawluck also created the dBASE name, in part because of potential trademark issues with Vulcan. The product name would be dBASE II to imply that it was more stable than the non-existent dBASE I. dBASE II was put on the market at $695 and the orders poured in.
The company was soon able to hire Ratliff and Long, and by early 1982 SPI had grown to 17 employees. By the end of the fiscal year in January 1982, had revenues of almost $3.7 million with an operating loss of $313 thousand. To further grow the company and instill sound business practices, the company hired David Cole as their second president (succeeding Tate) in February 1982.
As Cole began to take control of Ashton-Tate and the publishing business, Lashlee took responsibility for running the SPI distribution business. Among Cole's early acts were to hire an accountant to set up a financial system, install a management structure, and introduce processes to manage operations and orders. Cole's mission was "to shift the balance of power from those who understand how computers work to those who need what computers can do.
In 1982 dBASE II was ported to the IBM PC and the product was released in September of that year. With the growing popularity of ever larger hard drives on personal computers, dBASE II turned out to be a huge seller. For its time, dBASE was extremely advanced. It was one of the first multi-file products that ran on a microcomputer, and its programming environment allowed it to be used to build custom applications for almost any role. Although microcomputers had limited memory and storage at the time, dBASE nevertheless allowed a huge number of small-to-medium sized tasks to be automated.
October 1982 also saw the release of two unsuccessful products: Financial Planner and Bottom Line Strategist. Both were released at the same price as dBASE II, but were quickly discounted and eventually discontinued.
In May 1983 the company sold off the mail order and distribution businesses and officially changed the company name to Ashton-Tate. It also took steps to control its technology by creating an in-house development organization, and to diversify by funding two outside development teams: Forefront Corporation and Queue Associates. That summer, they released Friday!. By the time of the November 1983 IPO, the company had grown to 228 employees. The IPO raised $14 million. When the fiscal year ended in January 1984, revenues more than doubled to $43 million and net income jumped from $1.1 million (fiscal 1983) to $5.3 million.
In May 1984 the company announced, and in July shipped dBASE III as the successor to dBASE II. July also saw the release of Framework, an integrated office suite developed by Forefront Corporation and funded by Ashton-Tate. These were the company's first products released with copy protection schemes in an attempt to stop software piracy.
dBASE III was the first release written in the C programming language to make it easier to support and port to other platforms. To facilitate the rewrite, an automatic conversion program was used to convert the original Vulcan code from CP/M Z-80 and DOS 8088 assembly language code into C, which resulted in the beginnings of a difficult to maintain legacy code base that would haunt the company for many years to come. This also had the side effect of making the program run somewhat slower, which was of some concern when it first shipped. As newer machines came out the problem was erased through increased performance of the hardware, and the "problem" simply went away.
In fall 1984 the company had over 500 employees and was taking in $40 million a year in sales, the vast majority of it from dBASE or related utilities.
George Tate died of a heart attack at the age of 40 on August 10, 1984. David Cole became CEO briefly, but on October 29 announced his resignation and left for Ziff-Davis leaving Ed Esber Jr. to become CEO. Cole hired Esber because he was the marketing expert who launched VisiCalc and who built the first distribution channels for personal computer software. (VisiCalc was the first spreadsheet and is credited for sparking the personal computer revolution and was the first commercially successful personal computer software package.)
During Esber's seven year tenure, Ashton Tate had its most prosperous years and a few of its most controversial. It is also when Ashton Tate became one of the "Big Three" personal computer software companies who had weathered the early 1980s "shakeout", and was considered an equal of Microsoft and Lotus. Under his leadership Ashton Tate Sales grew over 600% from $40M to over $318M.
In November, shortly after Esber took over, dBASE III version 1.1 was released to correct some of the numerous bugs found in the 1.0 release. As soon as the 1.1 release shipped, development focus turned to the next version, internally referred to as dBASE III version 2.0. Among other things, the 2.0 release would have a new kernel for increased speed, and new functions to improve application development.
Esber's relationship with Wayne Ratliff however was tumultuous, and Ratliff quit several months later. Eventually a group of sales and marketing employees left to join Ratliff at Migent Corporation to compete with Ashton Tate. Later (January 1987), Ashton Tate would sue Migent for alleged misappropriation of trade secrets. Ratliff would eventually approach Esber about rejoining Ashton Tate and insisting on reporting directly to him. Jeb Long took over as dBASE's main architect in Ratliffe's absence.
In October 1985, the company released dBASE III Developer's Edition. Internally this release was known as version 1.2. It had some of the new features expected to be in the upcoming 2.0 release, including the new kernel and features primarily useful to application developers. 1.2 was one of, if not the most stable dBASE version that Ashton Tate ever released. Interestingly, it was also one of the least known and most often forgotten. Mostly, it was a release to appease developers waiting for 2.0 (dBASE III+).
In late 1985 the company moved its headquarters to the final location in Torrance at 20101 Hamilton Avenue in 1985. Development was spread throughout California, although dBASE development was centered at their Glendale offices.
dBASE had grown unwieldy over the years so Esber started a project under Mike Benson to re-architect dBASE for the new world of client servers. It was to be a complete rewrite, designed as the next generation dBASE.
dBASE was a complex product, and a thriving third-party industry sprung up to support it. A number of products were introduced to improve certain aspects of dBASE, both programming and day-to-day operations. As Ashton-Tate announced newer versions of dBASE, they would often decide to include some of the functionality provided by the third-parties as features of the base system. Predictably, sales of the third-party version would instantly stop, whether or not the new version of dBASE actually included that feature. After a number of such vaporware announcements, the third-party developers started becoming upset.
One particularly important addition to the lineup of third-party add-ons was the eventual release of dBASE compilers, which would take a dBASE project and compile it and link it into a stand-alone runnable program. This not only made the resulting project easy to distribute to end user, but it did not require dBASE to be installed on that machine. These compilers essentially replaced Ashton-Tate's own solution to this problem, a $395 per-machine "runtime" copy of dBASE, and thereby removed one source of A-T's income. The granddaddy of the compilers was Clipper. Eventually a number of these were developed into full-blown dBASE clones.
Esber was upset with the companies that cloned dBASE products, but was always supportive of the 3rd party developers who he viewed as an important part of the dBASE ecosystem. He did not believe nor support companies that cloned dBASE and leveraged the millions of dollars his shareholders paid to market. Starting with minor actions, he eventually went to great lengths to stop cloners with cease-and-desist letters and threats of legal action. At one industry conference he even stood up and threatened to sue anyone who made a dBASE clone, shouting "Make my day!. This sparked great debates about the ownership of computer languages and chants of "innovation not litigation".
As a result of this continued conflict, the third-party community slowly moved some of their small business customers away from dBASE. Fortunately for Ashton-Tate, large corporations were standardizing on dBASE.
Unfortunately, dBASE IV was both slow and very buggy. Bugs are not at all that surprising in a major product update, something that would normally be fixed with a "dot-one" release before too much damage was done. This situation had occurred with dBASE III for instance, and Ashton-Tate had quickly fixed the problems. However a number of issues conspired to make the dBASE IV release into a disaster.
Neither of these issues would, by themselves, kill the product. dBASE had an extremely large following and excellent name recognition. All that was needed was an update that addressed the problems. At the time of its release, there was a general consensus within Ashton-Tate that a bug-fix version would be released within six months of the 1.0 release. If that had happened, the loyal users might have been more accepting of the product.
Rather than do that, Ashton-Tate management instead turned their attention to the next generation of applications, code named Diamond. Diamond was to be a new, integrated product line capable of sharing large sets of data across applications. This effort had been underway for years and was already consuming many of the resources in the company's Glendale, Walnut Creek, and Los Gatos (Northern California Product Center) offices. However, once it became apparent that Diamond was years away from becoming a product, and with poor reviews and slipping sales of dBASE IV 1.0, Ashton-Tate returned focus to fixing dBASE IV.
It was almost two years before dBASE IV 1.1 finally shipped. During this time many customers took the opportunity to try out the legions of dBASE clones that had appeared recently, notably FoxBase and Clipper.
Sales of dBASE plummeted. The company had about 63% of the overall database market in 1988, and only 43% in 1989. In the final four quarters as a company, Ashton-Tate lost close to $40 million dollars. And in August 1989, the company laid off over 400 of its 1,800 employees. The Microsoft partnership also came to nothing as dBASE never worked well in this environment, and Microsoft eventually released Access in this role instead.
In 1990 he proposed a merger with Borland. During the first discussions, the board backed out and dismissed Esber thinking him crazy to entertain a merger with Borland, and on February 11, 1991, replaced him with William P. "Bill" Lyons. Bill Lyons had been hired to run the non-dBASE business and heretofore was unsuccessul. Lyons would ship dBASE IV 1.1, a product Esber managed and was virtually ready to go to Beta when let go.
After giving the board a merger package including individual bonuses of $250K and giving the management team repriced options and golden parachutes, the board and Lyons reinitiated discussions with Borland at a substantially reduced price and reduced joint oversight.
Wall Street liked the deal and Borland stock would reach new highs shortly before and after the merger. Some considered the $439 million in stock they paid to be too much.
The Borland merger was not a smooth one. Borland had been marketing the Paradox database specifically to compete with dBASE, and its programmers considered their system to be far superior to dBASE. The Paradox group was extremely upset whenever Philippe Kahn so much as mentioned dBASE, and an intense turf war broke out within the company. Borland was also developing a competitor product called The Borland dBase Compiler for Windows. This product was designed by Gregor Freund who led a small team developing this fast, object-oriented version of dBASE. It was when Borland showed the product to the Ashton-Tate team that they finally conceded that they had lost the battle for dBASE.
Nevertheless, Kahn was observant of the trends in the computer market, and decided that both products should be moved forward to become truly Microsoft Windows based. The OO-dBASE compiler was no more able to run under Windows than was dBASE IV, causing Borland to abandon both codebases in 1993 and spin up a new team to create a new product, eventually delivered as dBASE for Windows in 1994. Meanwhile Paradox was deliberately down-played in the developer market since dBASE was now the largest Borland product. Microsoft introduced Access in late '92, and took over almost all of the Windows database market. Further, in the summer of 1992 Microsoft had acquired Ohio based Fox Software, makers of the dBASE-like products FoxBASE+ and FoxPro. With Microsoft behind FoxPro, many dBASE and Clipper software developers would start working in FoxPro instead. By the time dBASE for Windows was finished the market hardly noticed. However, Borland then created 32 bit Visual dBASE and released it in 1997 to great accolades but a smaller market. (Visual dBASE has since been upgraded and is still sold as dBASE Plus).
When Borland eventually sold its Quattro Pro and Paradox products to Novell, where they would be joined with Word Perfect in an attempt to match Microsoft Office, Borland was left with InterBase, which Esber had purchased in the late 1980s and had its origins as a derivative of the RDB database work at DEC. Borland's ongoing strategy was to refocus its development tools on the corporate market with client-server applications, so Interbase fitted in as a low end tool and a good generic SQL database for prototyping. This proved to be the longest lasting and most positive part of the Ashton-Tate acquisition, ironic since it was almost an oversight and little known to Borland until then.
Overall, the Ashton-Tate purchase proved to unsuccessful. Several years later, Philip Kahn would leave Borland amidst declining performance
Any one of these would have been a surmountable problem, but combined they brought about the swift decline of the company.
Ashton-Tate's dependence on dBASE is understandable. It was one of the earliest killer applications in the CP/M world, along with WordStar and (on other platforms) VisiCalc, and was able to make the transition to the IBM PC to maintain its dominance. Its success alone is what created and sustained the company through the first nine years. However, the over reliance on dBASE for revenue had a catastrophic effect on the company when dBASE IV sales tanked.
In the end, the poor quality and extremely late release of dBASE IV drove existing customers away and kept new ones from accepting it. This loss of revenue for the cash cow was too much for the company to bear, and combined with management missteps, eventually lead to the sale to the upstart Borland International.
Framework was an integrated DOS-based office suite that combined a word processor, spreadsheet, mini-database application, outliner, charting tool, and a terminal program. Later versions also added e-mail support. Framework also had the distinction of being available in over 14 languages. Although DOS based, Framework supported a fully functional GUI based on character graphics (similar to Borland's OWL).
Framework eventually got locked into an industry battle, primarily with Lotus Symphony, and later with Microsoft Works. The market was never large to begin with, as most customers chose to purchase the large, monolithic versions of applications even if they never used the extra functionality. Borland later sold Framework to Selections & Functions, who continue to sell it today.
What was originally a deliberate attempt to copy the Wang's system now made the product seem hopelessly outdated, and it would require a major upgrade to remain useful. WordPerfect took advantage of these issues and took market share to a degree essentially lethal for MultiMate.
After the purchase was completed it became clear that the drawing product was inadequate. Although it was released as Draw Applause it never sold well.
The most recent release of Rapidfile was version 1.2 released in 1986, with versions available in several languages including English, French and Dutch. Although Rapidfile was created for the DOS operating system, information is available to show that it can be persuaded to work reasonably well in the DOS box of Microsoft Windows 95, 98, 2000 and XP, and also under GNU/Linux using the DOSemu emulation software.
As early as the winter of 1984, only a few months after the Mac's introduction, the company purchased a small Mac database developer and moved them to their Glendale development center to work on what would later be known as dBASE Mac. Soon after this, in early 1985, they agreed to fund development of a spreadsheet program being developed by Randy Wigginton, former project lead of MacWrite. Years later they added a "high-end" word processor from Ann Arbor Softworks, who were in the midst of a rather public debacle while trying to release FullWrite Professional which was now almost a year late.
Ed Esber and Apple Computer chairman John Sculley jointly announced Ashton Tate's family of Mac products in Palo Alto. dBASE Mac finally shipped in September 1987, but it was dBASE in name only. Users were dismayed to learn that in order to interact with their major investment in dBASE on the PC, their applications would have to be re-written from scratch. Adding to their frustration was the fact that it crashed a lot and was extremely slow. Given that the program was really a completely new Mac-only system, it had to compete with other Mac-only database systems like 4th Dimension, Helix and FileMaker.
FullWrite and Full Impact were released in 1988. Both were liked by reviewers and had leading edge features. FullWrite was an outstanding product, while Full Impact had the bad luck of being timed just after a major new release of Microsoft Excel and the release of Informix Wingz.
All three products were excellent at their core, but were really not viewed as a family and needed to link together more cleanly. They all also needed a solid follow-up release to address some of the bugs and performance issues. However, no major upgrades ever shipped for either FullWrite or dBASE Mac, and the only major upgrade to FullImpact shipped a full two years after release. Releases of Microsoft Word and Excel soon closed some of the feature gaps, and as the Mac OS changed the products became increasingly difficult to run. Microsoft embarked on a campaign in earnest to discredit and kill Ashton-Tate's products, at one point exaggerating the system requirements for FullWrite, and going so far as to delete Ashton Tate software from Mac dealers demonstration computers.
FullWrite was later sold off by Borland in 1994 to Akimbo Systems, but by that time Microsoft Word had taken over the entire market and they too eventually gave up on it. dBASE Mac was sold off in 1990 and re-released as nuBASE, but it was no more successful and was gone within a year. Full Impact simply disappeared.
A client-server database is a fundamentally different sort of system than a traditional single-user system like dBASE, and although they share many features in common, it is typically not a simple task to take an existing single-user product and turn it into a true client-server system. As the business world became increasingly networked, Ashton-Tate's system would become irrelevant without updating to the client server era.
Ed Esber and Bill Gates introduced SQL Server to the world in a joint New York press conference. The basic idea was to use SQL Server as a back-end and dBASE as the front-end, allowing the existing dBASE market to use their forms and programming knowledge on top of a SQL system. SQL Server was actually a product developed by Sybase corporation, which Microsoft had licensed. From a business perspective this had little direct effect on the company, at least in the short term.
dBASE continued to sell well, and the company eventually peaked at $318M in yearly sales. During this period, Esber hired some of the most brilliant database engineers in the industry, including Dr. Moshe Zloof from IBM, Harry Wong, and Mike Benson (who would later head Esber's efforts to rebuild a new dBASE).
Esber had previously decided to sue one of the clone companies involved, then known as Fox Software. By the time the case worked its way to court in 1990, Fox Software had released FoxPro and was busy increasing market share. If the court case was successful, Ashton-Tate could stop FoxPro and use the precedent to stop the other clones as well, allowing dBASE to regain a footing and recover from the dBASE IV incident.
These hopes came to an end when the case was thrown out of court. During the initial proceedings it was learned that dBASE's file format and language had been based on a mainframe product used at JPL, where Ratliff had been working when he first created Vulcan. The credibility of Ratliff was jeopardized by his alternate claims of ownership while at Ashton Tate and then supporting the roots at JPL after he left. All the facts were never sorted out and Ashton-Tate's competitors had a self-interest motivated field day in writing amicus briefs.
When the federal judge reviewed the work of his clerks he overturned his earlier ruling, and decided to hear the case on whether or not Ashton-Tate owned the language. Unfortunately, his earlier ruling had already done considerable damage. Eventually, as part of the merger with Borland, the U.S. Justice Department required Borland to not assert copyright claims in menu commands and the command language of dBASE.
Various articles describe the original mail-order company as being named Discount Software, while others call it Software Plus. (Both are correct, as they were different companies. Discount Software sold directly to consumers, while Software Plus was a distributor to VAR's and other resellers. George Tate also had a third company at the time he started Ashton-Tate, a small retail store chain called Softwaire Centre. That name would linger on for many years after the stores were no longer connected to George Tate and/or Ashton-Tate.)
Los Angeles advertising consultant Hal Pawluk named their first product dBASE II, as if it were a new and improved version. Hal also created a famous and controversial early dBASE magazine ad that read in part: "Those other databases are like bilge pumps. And there's one thing we know about bilge pumps - they all suck"
A possibly more accurate version of the infamous Ed Esber quote cited in the main article is: "Wayne Ratliff is no more important to the success of Ashton-Tate than the guy at the warehouse who loads the boxes onto the truck."
George Tate and a number of early Ashton-Tate employees were Scientologists.
In the early years of microcomputer software, any database product (such as dBASE II) that could link together two or more data files was often advertised to be "relational", even though by that time the term "relational database" had been given a much more rigorous definition in the academic world.
An early version of dBASE II included a license agreement that said the buyer had the right to use the software for 99 years. This was later cited by long-time PC Magazine columnist John Dvorak in a collection of software license oddities.
dBASE and InterBase continue to be developed and sold separately. dBASE Plus is available through dataBased Intelligence, Inc. (http://www.dbase.com). InterBase continues to be improved and sold by Borland (http://www.borland.com).
At one point in the late 1980s, the then president of Ashton-Tate (Esber was CEO) proclaimed that every product would need to bring in at least 10% of the companies revenue or be dropped. This statement had the immediate effect of limiting the company to 10 products or fewer. Further, with dBASE bringing in about 70% of the revenue, the other product managers quickly convinced the president to drop the requirement.