Memoirs by Primož Jakopin

On STEVE

2023

 

          This short story was initiated by a wish for a casual conversation about my program STEVE, the ST EVent Editor, by Steven Gregory of AtariCrypt, a retrogaming website dedicated to the Atari ST computer line.

 

About myself - before ATARI ST, the ST time, and onwards

          I was born in a linguistic family - father Franc was a professor of East Slavic languages, mainly Russian at the Faculty of Arts in Ljubljana while my mother Gitica was a translator. She translated 50 books, mostly fiction, from English, French, German and some other languages into the Slovenian. She also wrote poetry and fiction. Most of my computer-related pre-ATARI-ST time is described in the paper On microcomputers and on INES. My first contact with computer technology, however, was indirect and rather bizarre. In 1963 the first digital electronic computer in Slovenia, the Zuse Z-23, was installed in the building of the Slovenian Metallurgic Institute, next to the Nikola Tesla street where we lived. I passed it every day on my way to the primary school. At the side of the building was a set of trash cans, full of discarded perforated paper tape used by the teleprinter, one of the computer's input and output devices. Information on it was in form of holes - there were 5 rows of holes in the tape and every column (5-bit) represented a number or some other character. Children, including myself, used to play with it (I was 14 at the time). We even made a line of it which crossed the street about a meter and a half above ground, from one garden fence to the other, and it nearly knocked the neighbor we did not particularly liked off his bicycle. He did not notice it on time though it was in the height of his eyes.
 

          The ST time was my most hard-working, exhausting and close to bone-breaking time, but at the same time the most exhilarating, sublime and lucrative time, the horizons were opening, everything seemed possible. Those were the times when Europe, a little later than the US, wholly embraced the computer euphory, when the home computers leapfrogged from toy machines into something everybody could use to his advantage. The expectations, triggered by the first 8-bit home computers, started to be fulfilled. Atari ST with its 32-bit heart brought to everybody's desk not only a splendid, intelligent typewriter but also a true data-processing machine, capable of crunching numbers, text and even pictures. And I was in the eye of this hurricane. I could leave a job at the University Computing Center and devote all my time to the making of a universal tool, which found a great echo at home and abroad.
 

          The aftermath of my ATARI ST time was difficult at best. After enjoying great success in the European market after its 1985 premiere at the CeBIT computer expo, the world's largest international event of the time, ATARI ST home computer failed to unseat Apple MacIntosh as the best selling computer for the non-business sector of the US market. ST was a better machine, had more memory, a very sharp monochrome 12-inch monitor (Mac only had a 9-inch one) and a price tag of 3,000 German marks, versus 3,000 US dollars (nearly twice a much) for the Mac. But university buyers, students and teachers alike, were given a 50% discount by Apple in the US - none in Europe, and Apple Mac was an established brand with a classy design, a lot of software and peripherals, and the most charming presenter - Steve Jobs, all going for it. So in Germany most university users would buy ST's while in the US everybody thought - why buy an unknown machine if I can get the proven, stylish Mac for the same money. In 1986 ATARI sold 100,000 ST machines only in Germany (75% of all sales were in Europe), while the ST sales in the US just could not take off. Jack Tramiel cancelled any serious investment into the ST research and development in the following years and by 1990 it was clear that the ST is in a blind alley.
 

Transition to new hardware after 1989

          I was forced to look for an alternative computer platform where to develop EVA, STEVE's successor. The only viable one was the IBM PC and its derivatives. And here the sky was murky, full of dark clouds. To use the machine at its best, STEVE was not written in C or some other high-level programming language but in the assembly language of the machine's microprocessor, the 16/32-bit Motorola MC 68000. It gave STEVE a threefold speed advantage, extremely compact code, best possible use of machine's memory. The Intel 80286 microprocessor, used in the PC/AT, is just a 16-bit microprocessor, with 64 KB linear memory addressing capability versus 4 GB linear memory on the MC 68000. If you wanted to move 1 MB of text or data to another location, or just a little bit up or down in memory, it was a single command on MC 68000. And a set of commands on an Intel 80286. Like if someone gave you a task of moving a ton of water from one container to the other. On 80286 you would be given a bucket to do it, and on MC 68000 you could just pour the entire container into the other one. In one go. Not to mention memory swapping - programs, longer than 64 KB had to be segmented into smaller parts, loaded into memory for execution when needed. On a machine with 48 KB of available memory, such as the Sinclair ZX Spectrum, 16-bit processor was quite OK, but not on a 512 KB or 1 MB machine. 80286 with 14 16-bit registers was just no match for MC 68000 with 16 32-bit registers.
          It was such a step backwards that it could be compared to return from a computer text editor to a typewriter. Which eventually happened to me once, in the early summer of 1984. The major Slovenian home computer magazine, Moj mikro (My Micro(computer)) was about to start publication and its editor, Žiga Turk, approached me if I could contribute to the first issue. He had four articles in mind. I would gladly comply but I had other duty to fulfill - as a member of Army Reserve force I had a 14-day yearly military exercise in front of me. We settled for two articles: Can Slovenia penetrate the global computer market? and Microdrive units. The submission deadline was the same as the end of military exercise, so something extraordinary should happen to enable me make it. And it did happen. Senior officer of the exercise asked me to edit a short magazine that would be published towards the end, in the woods, reproduced on a photocopier in a small number of copies. The unit I belonged to moved around the country in the manner of the Slovenian resistance in WWII, sleeping in tents and using forests for cover, to avoid detection; it ended with a 30 km (20-mile) march, made in one day, fully loaded. I accepted the task and an old portable typewriter was assigned to me. Exercise during the day, editorial work in the extended evening hours, plus the two articles for Moj mikro. A few colleagues wrote maybe five articles for the magazine while the most difficult part, the editorial, fell on me. We all hated the army service, the entire Yugoslavian army was in Serbian hands, as was its language (not the reserve force) and we all thought that the fourteen days spent every year are pure waste of our time. To write an inspiring and enthusiastic editorial, promoting a totally lost cause, seemed beyond my capabilities. After two hours of brainstorming a solution popped up - what if I would try to play a role of a young Serbian officer, fresh from the military academy, who wants to prove himself in such a non-combat mission. I wrote that the scent of gunpowder should not be strange to us in these demanding times, that we should all be grateful for having such a splendid opportunity to enjoy nature at its best, to contribute to defense of our homeland marching through the scenic places, admired by the great travelers of the past since the Middle Ages, and the like. The commanding officer sincerely congratulated me, other feedback was less positive. When the soldiers of my small unit (I had a reserve rank of a corporal, they were mostly around 20, I was 35), not knowing who was the author (all the articles were unsigned), read the editorial, they came to an unanimous verdict: if they would only have one bullet in the rifle left, it would be spared to shoot the author. That way or another, it was all done, but never before (and later) was the writing so tedious, on typewriter the text had to be near-perfect in the first try.
          So Intel 80286 assembly language was out of the question, 80,000 lines of STEVE's MC 68000 code had to be converted to some high-level language. C would be preferable, but in the meantime C was also outdated, superseded by C++. While C was dubbed a structural programming language suitable for system applications and low-level programming applications, C++ is object-oriented, having additional features like encapsulation, data hiding, data abstraction, inheritance, polymorphism. Translated into my universe - if C code was three times slower than machine code, produced by the assembler, C++ was very likely a few times slower than C.
          Here another problem popped up. PC-compatible computers were using mainly DOS operating system, without any GUI, well into the 1990-s. In 1989, when I started the conversion of MC68000 code to C, Windows for the masses was non-existent, nowhere in sight and so I started working on EVA/DOS. There were no notebooks at the time, most PC-compatible machines of reasonable power were in extremely large and heavy full-tower casings, difficult to move around, let alone to smuggle it over the border. Legal ways to import such a machine into "communist" coutries such as Yugoslavia were still out of reach. So the distributor of STEVE for the German-speaking countries, Wolfgang Kieckbusch of Computer Technik Kieckbusch, Vielbach, a born salesman with a wide open mind, suggested a Compaq Portable 386 with 1 MB of RAM, built-in 40 MB hard drive and a 1.2 MB 5.25-inch floppy disk drive. The operating system was DOS, the machine sported a 10-inch amber gas-plasma display, 640 x 400 pixels, the same resolution as the Atari ST monitor, CGA. At 11 kg of weight it was, if not portable, at least transportable. With all possible discounts he could get as a computer dealer, it still costed 14,000 German marks (DEM), was by far the most expensive machine I ever bought.
          Getting the machine from Germany to Slovenia, smuggling it across several borders, was another matter. Export of high technology (i.e. computers costing more than 5,000 DEM) to communist countries was prohibited. There was a bus connecting Zurich in Switzerland with the Koversada Naturist Campsite in Istria, Croatia, Yugoslavia at the time. It was an overnight line, the bus departed from Zurich in the evening and arrived in Vrsar the next morning. On its way it passed Italy and entered Yugoslavia near Koper on Slovenian coast. My mother lived in Ankaran, a nearby village, while my aunt Sonja, her sister, lived in Switzerland, in Solothurn, less than one hour by train from Zurich. So Mr. Kieckbusch delivered the Compaq Portable 386 to my aunt, Germany and Switzerland were always well connected, and when my mother visited my aunt some time later, she returned taking the abovementioned bus line. Swiss/Italian border posed no problem, while on Italian/Yugoslavian border the Yugoslavian customs officer entered the bus very early in the morning, found a bunch of elderly ladies and gentlemen, mostly asleep, and did not bother to check their personal belongings - my mother had the machine on the floor, next to her feet. On the road junction between Trieste and Koper, where the road to Ankaran forks off, she got off the bus. My father was waiting for her and all I had to do was to pay them a visit. From Ljubljana where we lived to Ankaran there was a little more than one hour drive.
          I was also flirting with a UNIX version of Eva, and so I had both operating systems on the same machine, DOS and UNIX, and most of the 40 MB hard disk was used up. The installation of Unix from about 12 floppy disks took ages, and any mistake meant the whole procedure had to be repeated from the start.
          In two years the conversion of STEVE to C for DOS, certainly without graphical user interface, was mostly done. It also became clear to me that the Compaq, though more easily transportable than a tower-cased PC plus monitor and keyboard (their price also dropped significantly), was a very expensive choice. First true portables, with LCD screens started to appear and it was high time to get rid of the Compaq. Mr. Kieckbusch helped again, there were no buyers in Slovenia. In the meantime the country became independent, Western-oriented and to get the machine to Germany was no particular problem. But the sale was tough. In the end an owner of a concrete plant bought it. He was persuaded into acquisition by the argument that the new light notebooks are made of brittle plastic, while the Compaq is of solid iron and can withstand a kick into something hard. The price was also very reasonable - 2,000 DEM. In two years the machine lost 86% of its value. A small number of PCs followed, including one portable, the penultimate (and longer-lasting) being a mini-tower Intel-Core2-Duo-based machine which was written off as obsolete and given me as a present when I retired from my last full-time job (2000 - 2012). It lasted till August 2023 when I replaced it with the smallest computer I ever had, tiniest even than the Sinclair ZX Spectrum, a thirteenth generation of Intel's NUC machine.
          Both Apple MacIntosh and Atari ST were WIMP (Windows, Icons, Menus, Pull-down menus) based machines while Windows, such environment for the PC, was more or less an optional addition until Windows-95, which required an Intel 80486 processor and 8 MB of RAM for acceptable performance. In 1996 Programming Windows 95, a 1100-page handbook by Charles Petzold was published and it gave me two answers which I needed to make a Windows version of EVA - how to do it and why in C and not in C++. There were some low-level features in the GUI (Graphic User Interface) which were only available in C and not in C++, such as changing of the program menus and dialogues on the fly, directly from the program, one of the main advantages of STEVE. Examples in Petzold's book were written in C, not in C++, to make all the GUI features available. And if Petzold had to do it in C, so I reckoned, it will not be such a great sin if I do it in C, too.
 

How I obtained my first personal computer, the Sinclair ZX Spectrum.

          As already described, in early January 1983 it was clear to me that the Spectrum is the way to go. At the time the limit of a technical item value one could receive by mail from abroad in Yugoslavia was 10 (ten) US dollars. But my youngest brother Jernej (we were three, I was the eldest and the only one not engaged in J&J Design) was at the time working in Germany, in an architecture firm in Duisburg. He bought an 48 KB version of Spectrum and brought it when he came home by plane, hidden behind the waist, on the back side of the body. As he also wore a jacket, unbuttoned, all was fine, the computer (23 x 14,4 x 3 cm, 550 g) was flat enough that it could not be noticed as long as Jernej was upright, did not bend down. He passed the customs at the Ljubljana airport without problems. But he did not dare to bring the very bulky (800 g, 1.8 lbs) power supply, it was too big and heavy for a pocket and in the luggage it would be easily detected because of its metal content. So he sent it to me by mail and labeled the value as 9 USD. The guy at the post office customs was not stupid and recognized the trick. So he thought: Well, you shall have your power supply, but let me spoil the fun a little bit. And he threw it (unboxed) against some very hard surface, probably the floor. So I guess as I received it with the plastic casing cracked diagonally along its entire length. Yet the destiny had its way - the power supply had a broken casing but still worked flawlessly, for years, for as long as I had the machine.
 

STEVE in layman's terms

          In a few sentences: STEVE is a text editor with an own graphic user interface, designed to be as easy on your eyes as it goes, with many shortcuts which speed up the work. It has a data base functionality, data records are part of the file and can be edited, moved around, replaced just like any other text. For more demanding users there are additional facilities to be explored, especially for the lovers of language processing. It is easy to break text into sentences, words and letters, to further handle them as data items. Characters on screen and on printer can be user modified, the same applies to program messages, setups, menus. Useful for making a version of the program in language other than English, or to adapt it to one's particular needs. Two graphic editors for quick addition or handling of black and white images are included, a calculator, a DTP facility and a tool for building and playing Computer Assisted Instruction lessons.
          Before embarking on writing STEVE, I had 15 years of programming behind me, including two mainframe software packages, one designed and the other completed, operational and tested through a few years of use. IBIS, Integral Hospital Information System, included a novel idea of identifying hospital patient data records (or data about any other people in Slovenia) using an ID of name, family name, date of birth and an extra digit used only in case there would be two or more people with the same name and family name, born on the same day (none that I know of so far). To pack this info internally in just a few bytes the system would be based on a list of names and family names. BIOS (BIOmedicina Slovenica) was a program which handled the preparation of a data base with basic info on articles and books related to medicine, from which various indexes and reports would be generated on regular basis and which could also be searched. During the implementation I did not take a standard approach to write a separate program for every task, but I divided the workflow into a set of operations on the data base, which were controlled by a set of commands with arguments: sorting data records on specified data fields (keys), selection of data fields, searching on a given expression, formatting the ouput. So every task was just a short set of lines in this control language which made the entire procedure transparent, easy to put together and to change.
          During the making of BESS, my first text editor as the Sinclair ZX Spectrum had none, it was a natural extension to add such data handling operators to it, which I did in INES. The memory on Spectrum was very precious, however, and only the most basic operators were added. In STEVE this set was expanded by an order of magnitude, including two graphic editors.
 

My expectations with STEVE

          Frankly speaking, I did not have any special expectations with STEVE. I was just very very excited that was given a chance to build a piece of software that would be of most general use and where I would be able to show off all my ideas what such a tiny computing universe, revolving around the Motorola MC 68000 microprocessor, can do. Especially as I abandoned the idea of using ATARI ST's own GEM graphic user interface which I found limited (especially the tiny menu items in a single column box) and set off to write STEVE's own. So, for instance, the database menu of STEVE is nearly full screen big and the majority of database operators is available at once. The main functions, such as sorting and searching, are joined in the central column, while the "next best" routines are concentrated in the left column and the more exotic ones in the right-hand column. I also abandoned the GEM principle that the main menu has to be on top of screen at all times, taking away the most precious real estate, your screen space. On top and at the bottom of the screen I installed two narrow rope-like lines, which you touch with the mouse and the main menu appears. Even on PC, nowadays, there is no such luxury.
          From the feedback and reception of INES I guessed that STEVE will also find its place in the market and that I will be able to live off it. The presentation of STEVE at CeBIT Hannover in, I guess, 1987, confirmed these expectations. STEVE's lifespan was however connected to Atari ST's destiny which was, alas, limited to a few years.
 

The evolution from EVE to STEVE

          The path from EVE to STEVE, compared to moving from mainframe programming in structured high level language, mainly in STRUCTRAN (the paper about FORTRAN preprocessor is only available in Slovenian), to Z80 machine code, was a pretty smooth ride. Large memory, ample register set on the Motorola MC 68000 all contributed to rather comfortable working environment. I certainly had very much in mind the words of Professor Edsger Wybe Dijkstra, one of the fathers of structured programming and the man who elevated computer programming from a craft to a full-fledged science. When the micro-revolution started, he said: Microcomputers are not great. What he had in mind was that programmers, to squeeze all possible speed and power out of the tiny microprocessor, will have to forget about the elegance of structured programming in high-level languages and will be forced to move back to machine language. In the end STEVE was 80,000 lines in MC68000 language, certainly divided into a myriad of subroutines, but still a colossus hard to overlook and understand, especially after some time has passed. But to get the most out of Atari ST, to make STEVE faster and more compact than any possible competition, there was no other way to go.
          But if INES and EVE were just a herald of what is all possible on a micro, STEVE was very much like a real thing. Most of the ideas I had could be put into it.
 

STEVE, after completion, my feelings about it

          After completing STEVE, in July 1989, the time its English reference manual was published, I was completely exhausted but satisfied that the definite reference is finally complete and out. That the Dutch and Norwegian distributors (serving the Benelux and Scandinavian countries) will also have something to offer. But unlike a literary author, whose book sees the light of day and he knows he is done with it and can move on, software is like a child, it has its own life and the parents have, that way or another, to take care of it as long as it takes.
          The obvious continuation, to start a company in Germany to develop STEVE to the full and to support the sale and the majority of its users from up-close, fell out of consideration for several reasons. The Atari ST's cloudy future was one, followed by very poor chances that my wife would get a job in Germany, as a teacher of biology, plus we had two girls coming into the teen age. The language barrier was the main obstacle. It was just me who acquired a reasonable command of the language during the development of STEVE - most books, manuals for GEM and the machine, were in German. On the other hand I also lacked the entrepreneurial spirit, which my middle brother Japec, the driving force behind J&J Design, so amply inherited from our grandfather Anton (mother's side, he was called Zvonko). The latter was a master of salesmanship, he stemmed from a merchant family with a long tradition. A few years later, when I just re-booted my academic career, I was asked by my brethren to join the J&J Design team, to develop the velocity prediction program for boats, based on the hull and especially keel shape, the sails and other parameters, based on mathematical and fluid dynamics models. Such a program, ready made, costed 50,000 US dollars and was baked into an Apple MacIntosh, which came with the package. It was a most interesting challenge, but I could not accept it. It would take about two years to complete and my long-term academic prospects would be lost in the meantime.
          I gained a lot of experience making STEVE and would like to continue this work, to make an even better, more capable all-round program. To enlarge the character set from 8-bit to 16-bit, to add another, more general, dynamic data base model, to exploit the color capabilities of the forthcoming machines. From a developer's point-of-view itb was a normal way to proceed, from a commercial side a road to disaster. After the Atari ST's demise the sizeable pool of users, accustomed to STEVE, had to change the hardware platform to PC-compatibles. If I would deliver a PC version of STEVE, plain and simple, in a year or two, it would be a big success. But I could not. My heart would break.
           There were other considerations that had to be addressed. The burnout syndrome for instance. Ten years, since the 1982/83 New Year's Eve, I was working around the clock, most weekends, most holidays. It was even worse as I was getting desperate because no matter how hard I worked, how many hours I spent in front of the screen, the target of completing the work, the finish line was nowhere in sight. I always had the impression that I should do a little better. As the spring was coming every year, the finest time of the year, so was the CeBIT computer fair in Hannover (in March), where new version of STEVE had to be presented, with new and even more exciting features. Often finished just hours before the fair. There are 900 kilometers (560 miles) by car from Ljubljana to Vielbach in Germany, the starting point to Hannover fair. On one occasion I planned to depart in the morning but this and that still needed to be fixed in STEVE and I departed at one o'clock in the afternoon. In the evening I still had about 200 km to cover and I was so tired already that I would see two roads instead of one. When I started to see three I knew that I will fall asleep at any moment. I stopped for a 10 minute-nap and continued at a slower pace afterwards. In Hannover a growing pain developed in my upper jaw and after I returned home I visited a few doctors. The final diagnosis was: The jaw is fine, all the teeth are in place and in good condition, you are just overstressed.
           I needed a break. And I got one, working on the transition from STEVE to EVA on PC, later joined by NEVA (server side EVA search engine, furnishing answers to queries transmitted by Internet browser clients through CGI interface), at a slower, manageable pace. All the work assignments I faced in my later academic years I solved either through existing text and data handling routines in EVA, or by adding to their set. EVA grew to 156,000 lines of code (September 2023) and with it I was always able to make a fairly accurate estimate of how much time a certain task would take. And if I multiplied that time by two or three, which I usually didi and which would give me so cherished breathing space, it was still very much acceptable to my superiors. Programming every task separately would take much much more time. And people who stick to the deadlines they set in advance, are always at great advantage to those who are always behing schedule. STEVE made my later life easier and more enjoyable.
 


 
Primož Jakopin: Circles. Digital Color Composition, 2013 & 2015.
 

STEVE in Slovenian and the consequences

          INES had the main market in my country, Slovenia. The majority of users was here, I had a certain status because of it in my homeland and the potential of INES's successor, STEVE, was immediately recognized by Mladinska knjiga (Youth book in English), the main publishing house. Mladinska knjiga (MK) was at the time also trying to diversify its scope to other areas, such as personal computers which were quickly becoming the main writing tool. MK was eager to support me in all possible ways, including giving me the opportunity to present the program at CeBIT, the main European computer expo, in March 1987. So it was only natural that I wrote the first STEVE manual in Slovenian (Steve : urejevalnik besedil za Atari ST, 247 pages), published by MK at the end of 1986 (STEVE version 2.0), they also promptly arranged a translation into Serbian/Croatian (Steve: uređivač tekstova i jednostavnih zbirki podataka, again 247 pages), also published by MK in 1987. This was the language of the majority of Yugoslavian nations, Slovenians (2 million) were only one tenth of the population, but Slovenia was closest to Europe and the most developed part of the country. The reception STEVE received at CeBIT in March 1987 was very positive. The director of Mladinska knjiga export department and myself had a meeting with ATARI USA representatives, about delivery of STEVE, in English, with every ST computer sold. It would add very little to the price of the machine, but all in all, because of the volume sold, it would be big money. Yet the MK department director wanted one half of what ATARI would pay for STEVE. The other half would go to me. I had no experience in marketing, I just saw my great effort in putting the program together, 80,000 lines of code in machine language, based on two years of INES and all the sacrifices it costed me and my family (the title of the INES review in the leading Slovenian home computer magazine Moj mikro/My micro(computer) was One hundred Saturdays and one hundred Sundays for INES) versus a few hours, at best, of work for the MK director, and making both efforts equal on the scale was just too much for me. I was foolish enough to refuse the offer under such conditions. Yet the whole thing was a clear signal to me that a reference manual in English would be a very good idea. The Slovenian manual was just of no use in Hannover.
          As Professor Władysław Turski of Warsaw used to say, if you have something important to communicate to the world, write it in English. If you do not, somebody else will have to do it and his translation would inevitably be worse than what you would produce. And if what you want to tell the world is not important, why bother, you better stay silent.
          I should have done the same as I did with the INES reference manual. I wrote it in English and my colleague, Janez Kanič, translated it into Slovenian. By the summer 1987 I wrote a provisional reference manual in English (STEVE version 2.3). A proper, complete reference manual takes time. A year was not enough, especially as I was also developing and expanding STEVE in the meantime. Definitive book, STEVE reference manual : text, graphics, data base, DTP and CAI on Atari ST, 608 pages, was (self-)published in July 1989 (STEVE version 3.29). The number of copies printed was close to the number of pages, 605. The main market for the program was Germany and at the end of 1987 the first translation of the provisional manual was translated into German, by Günther Weber, an employee of the German distributor, Computer Technik Kieckbusch GmbH (STEVE: Deus ex machina - title of the report in German computer magazine ST Computer). But the translator was a computer salesman, and a better translation was called for. A year later, not without hassle (I virtually locked the two authors into my cottage south of Ljubljana for a week to force them to finally complete the book), a proper German manual came to light, written by Peter Wieser, versed in computers and Klaus Detlef Olof, a professor at the Faculty of Arts, both from Carinthia.
          The difference between earnings from STEVE, as it actually happened - a small car and a small cottage, and as it would be, should the English and German manuals be published a year or two earlier (a decent car and a proper house)? Considerable.
 

Interest from the ST community, at its prime time and now

          The echo from the ST community, after STEVE was released, before all the sales figures, showed me that the ideas I implemented in the making of the program, my vision of such a tool, were correct. STEVE had its own graphical user interface which made a difference from other programs right from the start. It also sported white letters on black screen by default, again unusual for a text editor. CRT screen of the Atari ST monitor emitted electrons on white parts of the screen so 8% of the screen was emitting harmful rays instead of 92% if black text would be displayed on white background. The letters were large, typically 64 per line, 22 lines per screen. Yet it all contributed to work which was easy on the user's eyes, even more so if the user was afraid to lose sight gradually due to long hours at the computer screen. That the choice was correct is attested by the fact that the author continues to work at the computer screen several hours a day, without glasses, though he was born in 1949. Most of his writing is still done in the same fashion as the users of STEVE - he works with EVA, STEVE's successor that inherited the same screen layout.
          Most of the feedback was directed to the STEVE distributor, and reached me indirectly, in person mostly at the CeBIT computer fair in Hannover, every year in March. There were some fans with particular needs or desires for changes in the program, they regularly showed at CeBIT. The relationship was amicable, a user from Heilbronn would occasionally visit me in Slovenia, another from Schiffdorf I visited at his home.
          A different story were the users from my close vicinity, from Ljubljana for instance. Most were institutional, from libraries, documentation centers, archives, from the Institutes of the Academy of Sciences. Their needs and wishes were taken into account promptly, some remained the users of EVA and NEVA, STEVE's successors. From time to time, even as of 2023, I receive a bug report or some other comment.
 

The usage of STEVE

          The vast majority of users were, very probably, people who wanted to try something new, there were a lot of text editors to choose from. They got used to the way STEVE worked, and liked the additional capabilities the program offered. The transition of a bunch of data, written as text lines, to a simple database with data records, composed of textual and numeric data fields, separated by the lifted dot (·), data item separator, was in STEVE completely natural and very smooth. Sorting, searching, mailing list, breaking of text into sentences and words, character statistics, were all at one's fingertips, without leaving the text editor. This synergy is probably the greatest appeal of STEVE, its predecessor INES and the followers EVA and NEVA.
          On corporate level, many librarians found good use of STEVE in their daily work, as at the time the centralised book and article catalog of all public libraries in Slovenia was in design stage. Around 1991 libraries of all the departments of the Faculty of Arts at the University of Ljubljana, some 15 of them, bought Atari STs to start their automatisation with STEVE. And certainly all needed maintanance and adaptations of STEVE. In 1992 my freelance inventor status (self-employment) which I had since about 1988 was no longer tenable. The income from STEVE sales in Germany followed the decline of Atari ST sales and was dwindling for quite some time while the servicing the increasing number of domestic corporate users was consuming more and more time, which I needed to develop EVA, the successor of STEVE.
          The manager of the Faculty of Arts, Amalija Šiftar, saw that the individual contracts for STEVE maintenance and courses of its use for all the libraries would simply cost too much and made me an offer I could not refuse - a post in the emerging Faculty Computing Center, where I would be half time involved in library software support and half time as a lecturer of an optional subject Computer science for philologists. The latter would lead to a PhD and possible full-time teaching tenure, starting September 1992. New recruitments at the faculty were halted due to cost savings at the university level but Mrs. Šiftar found a bypass - I was paid from the funds, allocated to the central heating. In hope that the winter will not be long and very cold. Her offer saved my day and provided a career path in language technology for the next 20 years of my professional life.
 

My owning of Atari ST. Echo from STEVE users after so many years.

          None of my early computers is with me now. Not even the Spectrum. I do not remember how I obtained my first Atari ST computer, but I do remember that I sold it when the STEVE story was not yet complete. All changes to the program that I have made afterwards were done on an Atari TT which I borrowed from Matija Gogala, a Slovenian entomologist, who used it for his research in insect bioacoustics, sounds, often resembling music, the beetles make.
          After all this time, it is already 30 years now, I still hear from an occasional user, who started with STEVE but is now using Eva, either for his/her daily work of for solving some other problem, for which he knows the program would be more suitable than any other software, or combinations of software or which would require own programming. Or who discovered either a flaw which needs to be fixed or just wants to know the shortest set of commands which would save him time in getting the work done. There are always several ways to solve a problem and it is a great difference if you need 10 steps to get it done instead of 100. The feeling after finding an elegant solution to a problem is just infinitely better.
 



This paper in Slovenian
 


 


Related posts:
 


 


Page written and posted by Primož Jakopin; send inquiries and comments to primoz jakopin guest arnes si (insert dots and at sign as appropriate). Contributors to STEVE, as of 1989: Andrej Vučkovič, Jure Dimec, Matija Gogala and Boris Kryštufek, Gitica Jakopin, Andrej and Maja Kranjc, Milan Orožen-Adamič, Tihomil Šlenc and Philip J. Burt. The paper was started on May Day 2023, finished in September 2023; last update: 22 October 2023.

URL: https://www.jakopin.net/papers/memoirs/On_STEVE_en.php
    963