01/04/2025
Teppo Määttänen, Finland
"My Progress Story" - part I
Foreword
This presentation does not aim to be a comprehensive story of Progress Software's operations, not even in Finland, but a personal reminder. The text contains personal snapshots and assessments of the stages of the activity. Part II is designed to include my years as a Progress Software employee, Part III as an independent software vendor and consultant.
Introduction
“Progress”, which is discussed in this presentation, is an application development environment for making “computer applications” for businesses and public administration. It is a set of software products that includes a database, its own programming language, and other tools needed to develop, deploy and use applications. Today, the name of the product is “Progress OpenEdge” or “OpenEdge”.
The original developers of the product, of which there were three - Charles “Chip” Ziering, Clyde Kessel and Mary Szekely, had a vision and plan for development tools that would form a comprehensive whole. You need to find in the same package everything you need to create and use applications. All aspects are produced by the same design organization and work seamlessly together. This company is also responsible for the flawless functionality, training and technical support of the software and receives user ideas for further product development.
I came across Progress myself for the first time in my hometown - Turku. After an internship in a programming course - in the summer of 1985 - I took a two-week advanced course, and as a new subject, our teacher Jouni Kokkonen started teaching us the syntax of unix operating system and Progress development. These were both new things whose concepts were at least partially confused. The exercises were performed in company "Nokia Data" in Kupittaa, Turku, on a computer called Fortune Miniframe.
The database was created with the command:
prodb demo demo.
The database programming started with the command:
pro demo.
English sentences were written in a text area, which Jouni called “command mode”. We learned then that it was a Progress editor and English was the Progress Forth Generation Language (4GL).
At the time, I couldn’t imagine that in another 35 years I would be dealing with this tenacious programming technique!
However, this was not the actual beginning of my Progress career, but it had to wait until 1987, since when I have been using the product continuously for over 30 years - daily.
Nokia Data
Later, Heikki Leinonen has explained how Progress was introduced in the Nokia Group - one of the first Finnish Progress customers. Later, focusing on mobile phones and networks, Nokia became the largest Finnish company, the global mobile phone product leader.
Representatives of Nokia's top management visited the Comdex trade fair in Las Vegas, where Progress had a stand. The company name at the time was ”Data Language Corporation”, and product version 1 was not commercially available. But willing trade visitors were given a patchwork floppy. At Nokia, the floppy was delivered to Nokia Data in Turku, and - among other - Heikki Leinonen began to make applications with it, first the chart of accounts for businesses.
Nokia Data's problem at the time was the obsolescence of the software hardware platform, as it had been decided to stop manufacturing its own ”Mikko” computers. The gaze was focused on the unix and Progress was also focused on it. Shortly afterwards, Nokia Data decided to reform the development of enterprise systems and do so with Progress.
Nokia's corporate management later made a similar decision: in a closed ballot, Progress got all the votes and Oracle was second. (Since then, Nokia has focused on mobile phones and networking devices and sold everything else in its industry, including the manufacturing of computers and software. Many of these companies became long-term Progress customers.)
Basic ideas for Progress products
The ideas for Progress software were not born out of nowhere, but the developers had extensive experience in developing business applications. These applications consist - to put it simply - of a database and application software that is “coded” in some programming language. In the 1980s, the most important programming language for business applications was COBOL. The database was separate software that was attached to the application using complex subroutine technology - depending on the database. Computer applications were used most in large companies because the hardware and software - that is, software development work - were immensely expensive.
In the 80s, the so-called minicomputers were capable of making smaller systems. Business systems needed the so-called mainframes.
Business applications are high-quality software. They must be used by many users at the same time, ie duplicate data changes must be prevented, but as subtly as possible. Database changes must be treated as transactions, ie implemented as a whole or canceled. Systems need to recover from potential failure situations such as power outages or telecommunications outages.
In general, these requirements were met only in a mainframe environment, and even there with complex programming and the use of expensive add-ons.
Another challenge was the general inconvenience of programming. Not so much how an operation is performed in a programming language, but how to prepare for all possible errors and disturbances.
The founders of Progress created a vision that was the ideal programming language for an application developer to cover database processing, application business logic and interface management.
At this time, the most common user interface was the so-called terminal. The terminal connected to the mainframe consisted of a display, typically with space for 24 rows and 80 columns of characters, and a keyboard.
In order to achieve the interactivity of the user interface, Progress introduced the so-called full-duplex displays, where each keystroke was echoed on the mainframe before it appeared on the display. This allowed the program to be able to respond to each key, for example a function key (F1, F2, etc.). There was no mouse or graphics.
Client-Server
Although the platform for the first applications was a host computer (now called a server), the architecture was made two-tiered, i.e. client-server, from the beginning. The application program was run in a client process, the server process managed the database and executed data commands from the client.
"Virtual Machine"
Programming languages are typically divided into interpretable and compilable languages. The compiled program, i.e. machine language, is executed as such in the computer's processor. Interpretable programs run a machine language interpreter program, but the application program is untranslated source code. The problem with interpretable programs is that syntactic errors in the language are only revealed at runtime when the ex*****on of the program encounters an incorrect point. (One of the most widely used programming languages today is JavaScript. According to Brian Preece, JavaScript is the second worst programming language ever. The worst is BASIC.)
Progress ended up with a kind of hybrid solution. The program file is compiled, but not into machine language, but into a kind of device-independent intermediate format executed by the interpreter. The compiled program is called R-code on the basis that the compiler produces a file with the same name from the source language file, but appends “.r” to the ending. The source language program file can be of any name, but it is common practice to enclose it in “.p”. The abbreviation p comes from the word "procedure" and the abbreviation r from the word "run". This interpreter - like the compiler - is a Progress Client and corresponds to the current technical term “virtual machine”.
With the proliferation of personal computers — PCs running rudimentary MS/DOS — from the mid-1980s, the same Progress applications were made to work on the desktop relatively easily. The database was placed on the server where the server processes were started. Progress had been developed to work with various network protocols to implement this database connection.
Progress at OSVA
I worked for Oy Saab-Valmet Ab ("OSVA") - Finland's only passenger car industry - as a software designer from 1985 until the end of the decade. The company had a fairly large computer department, which was divided into an operations department (operators, equipment maintenance, etc.) and two system design units: materials & economy and production control & manufacturing. I myself work in the latter team in the first phase of production planning development tasks.
The design of the cars was done mainly in Sweden. The hardware architecture in the company was ”two-tiered”, with product constructions and plans created on a mainframe computer that was Sperry/Univac. The finished production plans were controlled at the factory on another computer, which was during my time first Ericsson, but was later replaced by a fault-tolerant Tandem computer. All had COBOL as the programming language, IDMS a network database in Sperry and Tandem-SQL in Tandem.
What was significant for my own career development was that Sperry also used an 4GL application tool "MAPPER", which was used to make smaller stand-alone systems. The main developer was external consultant Tuomo Nyyssönen, who programmed with both COBOL and MAPPER. Together with the production designers, he made a construction system maintained at OSVA, which included products manufactured in Finland. The main database was operated by the Swedes and operated in Sperry and did not allow for the flexibility required by the design. At some point, I was able to maintain these MAPPER applications, which had also been made for the human resources department, e.g. payroll.
Crucial to Progress was the introduction of Intime Open software in 1987 for OSVA's financial administration. At the same time, the architecture choice for our own application development was underway, and TCP/IP was found in network traffic. In application development, it was also decided to try unix machines, for which, of course, the development method had to be solved. With the Intime Open, NCR unix device and a Progress database came to the house.
In the NCR unix Progress development methods were assessed. The alternatives, of course, were Oracle and Ingres was also thoroughly researched. Despite their promises, the gurus of the State Computer Center could not get Ingres installed on the most powerful DOS machine on the market, so only Oracle and Progress could be compared. Training courses were held for both, learning both Oracle’s SQL language and Progress-4GL.
Intime Open was acquired from Progress vendor Unic Oy, but the hardware was supplied by Mercantile Data. It was NCR Unix, and Mercantile Data also supplied the Progress products for the app. I got into Unix training with three developers, and Jorma Kohonen held a basic Progress course at OSVA even before the development tool had been selected. He also did a lot of developing Intime subscriptions, and the Progress code began to emerge.
One afternoon when I left for home, I asked IT Manager Ilkka Laitinen which one was chosen. He briefly stated, "Oracle was too expensive."
After the training, pilot applications had to be made with Progress, but only the demo application that came with the product was available. Jorma asked if his friend could come to help develop the pilots, and one morning Ilkka Pärssinen came to our booth by train with his backpack. He brought with him the “Kilvas Frames,” which became my first Progress Application Framework. Jorma and Ilkka then had the company Kilvas Oy.
The technology of the Kilvas frames was based on standard include files, which made the different application modules similar in terms of user interface and database handling. The new module was created by copying about ten files to a subdirectory named table, editing them one at a time with the sed editor, and making the necessary field lists on the table and form screens. The basic "CRUD" operations were complete. Productivity was at least 10 times better than in COBOL development carried out at the same time.
At the same time, personal computers came to desktops. MS-DOS devices were used for word processing and spreadsheets, but also for making their own applications. Pascal language was encoded in email, and enterprise applications were made using DBASE / Clipper technology.
With the help of Ilkka, all MAPPER applications were gradually ported to the Progress environment, of course further developed. As a new application, e.g. a system in which the staffing needs of the assembly workstations were planned. When he came to work, it was seen whether there were absences and the need to use deputies.
Finland's first Progress users
Progress came to Finland unlike other Nordic countries. At the beginning, company called Modern Software Europe was responsible for resale in Nordic Europe, in Sweden, Norway and Denmark. Progress bought the entire company in 1987, when it had run into considerable financial difficulties with its massive marketing, and Progress was the largest creditor. MSE's offices became the first subsidiaries of the corporation, and Progress received a very international stamp from the beginning.
It was decided to apply for a resale channel in software houses in Finland, and more were found. Vaasoft Oy, based in Vaasa, found the founders of Progress at the Comdex trade fair, which presented version 1 - which was not commercialized. Vaasoft built complete systems for companies with Progress as early as 1984. Valmet Automation, based in Tampere, built paper machine reporting with the first version. The above-mentioned Jorma Kohonen and Ilkka Pärssinen are the ones in these projects. Nokia Data made significant applications for the Nokia Group.
Progress was not satisfied with software houses in Finland alone, as they focused mainly on their own product development. Gary Johnson told me that Nokia Data, for example, was trying to embed Progress into its own developer and its own applications and didn’t want to offer the product to other software houses at all.
Mercantile Data was introduced as a new channel. Progress was initially included in the company's sales of unix equipment, at which point Mercantile sold complete packages: the NCR unix computer and with it the Progress database. NCR equipment salespeople handled Progress sales under Wiesek Sokolowski, and Jorma Kohonen was responsible for everything else. Everything else included software package shipments, technical support, product training and consulting.
Sales progressed rapidly, especially to software houses, and in the same week in 1986, Progress was sold to Unic Oy in Helsinki and Systla Oy in Lahti. Their applications then became Finland's leading Progress software for years, and they are - still developed - in some positions today. Software owners have changed, but current products still have the original Progress 4GL code.
As operations expanded, Mercantile Data established its own department “System Services” for Progress operations and was led by Jyrki Maukonen.
European User Conference
In the autumn of 1987, Progress organized the second user meeting in Vienna in its history. The first was held the previous year in Amsterdam. I was very impressed to meet the American staff at Progress who gave great performances. In particular, Chip Ziering’s self-confidence and Chad Carpenter’s patience in dealing with user questions and suggestions came to mind.
SQL?
I myself asked Chip Ziering later at the Paris User Conference in the fall of 1990 why Progress did not choose SQL as its database language. He was the chief designer of the product and the first evangelist. The answer was clear; SQL was poor in online transaction processing (OLTP).
The goal is Progress was to create a universal application development language, and SQL was completely inadequate for that. Compared to SQL search statements in Progress FOR EACH, FIND, etc., the record fields were immediately processed in the “record buffer”.
The name of the company behind Progress at the time was descriptively Data Language Corporation, as the goal in Progress was to create a programming language suitable for transaction processing, in which database searches and transactions took precedence.
One of the advanced features is the indexes of the tables: in SQL, the index could only move in one direction, while Progress had FIND PREV and other very versatile search phrases already in the next versions.
Air travel
It was a memorable trip to Vienna and back. I heard that Peter Tähtinen is making a trip on a small plane, and it is possible to take part in it. The group from Jyväskylä picked me up from Turku Airport, from where we flew directly to Vienna after the Malmö stopover. We flew over the GDR and saw from below the beautiful cities of Prague and Vienna. The return flight over dark Europe seemed between illuminated Berlin and the stormy landing in Turku was an experience. The group included shy student trainee Veli-Matti Korhonen, Progress colleague for many years to come.
The following year, I participated in a similar conference in Stockholm, where I discussed e.g. with CEO Joe Alsop.
Master Progress course
The first Progress advanced course was held in Helsinki in the spring of 1989. Steven Feinstein from Boston was a teacher. Since then, he became one of my most important colleagues when I became a Progress employee. He was a great performer and later the head of the support organization for all our Pre-Sales organizations at headquarters. In Helsinki, I had interesting conversations with Steven. He was convinced that in a few years Progress would overtake Oracle and other similar relational databases in the market.
In connection with the course, a user meeting was also held, where a first attempt was made to establish a user group. When no suitable leader was found, the project was postponed in the 1990s.
Parma Oy
At the turn of the decade, our family moved to Forssa a small city in Southern Finland. I had already applied for a job by visiting the city companies in the autumn of -89. The largest employer in Forssa was Parma Oy, founded by Armas Puolimatka. HR Manager Markku Saviniemi was excited about my stories on the car factory's advanced computer systems and directed me to the speeches of the people in charge of the Wood Industry. The title of the job became the window factory “Operations Instructor”, which also meant the development of an operations control system.
The systems in the construction industry were rudimentary compared to the car factory. The IBM S-36 used an Avonius application programmed with RPG. It was mainly a system for depositing and invoicing orders, and was paralleled by the actual operation of quotation accounting, product design, etc. All systems were separate from each other. The bid calculation used a Pascal-programmed file processing on a PC without a database. The design included CAD workstations, each with its own storage and printing system. For the S-36 mainframe, the secretary tapped the orders at the latest when they were ready products. The transition from Saab-Valmet's CIM environment was like a return to the Stone Age. (Later I met Erkki Avonius in Tallinn when he had large software businesses using Progress!)
Of course, I had in mind an integrated system that covers all stages of operation. At OSVA, the systems had been done as their own work, but the option was to purchase a ready-made system.
We made visits to Systla Oy in Lahti and got acquainted with their Progress-based "Toolkit" ERP system. At the same time, I was trying to make a prototype into an operations control system. The company purchased information on new construction projects from Rakennusfakta and submitted the data to the application as files (crumbs). Because the app was difficult for end users to manage and lacked search capabilities, I made a new app for the company’s vendors using Kilvas frames framework.
The Parma Concrete Industry was inquired about my ability to make them a production maintenance and material storage system. I had some experience with the system made at OSVA Tandem and got acquainted with similar systems. In this sense, the project was started at the same time as the economic recession of the 90s made it possible. The condition was to establish our own company, when we established TR-Tiimi Oy with Risto Laatu in the autumn of 1991. Risto began to offer our maintenance system to other companies, but the results remained modest.
Establishment of Progress Software Oy
Progress's globalization strategy was to establish distribution channels in all possible countries. When the resale channel grows functionally sufficiently, a wholly owned subsidiary will be established in its place.
In Finland, the establishment of its own subsidiary was announced for the first time in public at the European User Conference in Paris. At that time, Dave Vesty, who was responsible for international sales, and Rolf Carlstens, a Swedish subsidiary, presented a plan for PROGRESS SOFTWARE OY, to which Mercantile's operations were transferred in their entirety. Jyrki Maukonen, who was present, remained silent, even though he had already received confidence in the new CEO. At the user meeting in Paris, I represented Parma Oy together with Pekka Rantanen.
The opening of the office on Sinikalliontie in Espoo was celebrated in the spring of 1991, where I took the first product order from TR-Tiimi.
In the fall, additional reinforcements were sought for technical support and training, at which point I thought I would seize the opportunity. The recession in the construction industry had been affecting for some time now, and I thought of working for Progress despite a 106- kilometers commute for a few transition years. This period eventually stretched to 13 years.
