Damian Scattergood

Writing a complete computer game with a beginning, a middle and an end is very difficult. You spend ages learning how to do it. You have to study the electronics, to buy every programming magazine you can find, to identify people who can help you and make phone calls to them. My office was the local phone box or, sometimes, the one person on the road who had a telephone in their house.

The first game I sold was a maze game called Insomnia for the ZX Spectrum. I was 15 or 16 when it came out in 1982. I founded my first company, Scatz Computer Games, when I was just 17.

From then on I kept looking for a big break in the video game business. I got to know the Spectrum backwards. If there was a spare byte somewhere on the machine that I could use, I used it. In 1988 I started going to the Irish Amateur Computer Club, which met every second or third Saturday in Powers Hotel. I went to game exhibitions in the UK and did everything I could to talk to people over there. I also became very good at dumpster diving to get more hardware.

My first job was with Mentor Educational Services, writing educational courseware for PCs and home computers. I was their only programmer and we had a French translator to help me localize the software. After that I worked for two game developers – New Concepts in Carlow and Emerald Software in Waterford. At Emerald we had a great success with Vigilante, an arcade game that we converted for Spectrum and Amstrad. It got into the sales charts and onto the front page of games magazines.

In 1987 I took an AnCO training course in microcomputer management. This included computing classes which, because I was already a programmer, I did not want to take. As an alternative, AnCO gave me access to a North Star Dimension and I taught myself how to write applications in the C language.

Commology was looking for a C programmer who could also do machine code and I took the job. The company had contracts to write software for handheld Epson machines used by poll tax collectors in the UK and for an alarm system to prevent fires in grain silos. I also worked for a while at Datatech International, which distributed AutoCAD products and developed its own CAD tools.

What I really wanted, however, was to build a business of my own. In 1992 I set up my second company, Scatz Computer Graphics, and joined a start-your-own-business course at FÁS. I still ran Scatz Computer Games in the background as well.

Scatz Computer Graphics consisted of myself, Derek White who was an architectural engineer and Gary Rafter, a mechanical engineer. We hired ourselves out for custom development work. One of those jobs was helping Symantec to develop Norton Commander.

Symantec had established a software development group in Blanchardstown in 1992 – based in a building next door to the European packaging and distribution operation that it had opened in the previous year. The company was expanding fast through acquisitions and had bought Peter Norton Computing in 1990. The software group in Blanchardstown hired us to work on a new international release of Norton Commander file manager for DOS users. More specifically it wanted a translated version for Germany where the product had been a big seller. Tony O’Dowd was our project manager.

Very soon this work for Symantec became my full-time job. In 1993 Symantec hired Derek, Gary and myself and we closed Scatz Computer Graphics. I was employee number seven in the software development group.

The main managers there were Tony O’Dowd, James Grealis, John Rowley and Tommy McShane, all led by our VP Stephen Brennan.

This team produced localized versions of Timeline, Symantec’s project management application, in 1994. Pictured in front of the company’s premises in Blanchardstown are (l-r): Peter Heffernan, Stephen Roantree, Karen O’Connor, Damian Scattergood, Mark Roantree and Eric Schyberg, who was based in Novato, California.
Photograph by Damian Scattergood.

After we produced a couple more DOS releases of Norton Commander I became development manager for a Windows version. This work was done in Ireland because Commander was seen as a dead product in the US. I was responsible for both the software development and the software localization teams. Symantec released Norton Commander for Windows in 1996. By then we were shipping packages in thirteen languages. Translation and localization had become central to my work.

Symantec did not allow any randomness in its localization process. The way that we did things was very rigid.

This process began by testing early versions of a new product on foreign language computers, including machines that displayed double-byte characters. Then we performed a pseudo-translation of the application into each of the languages and liaised with the development team in the US to sort out the bugs. It was only when a new product reached the beta stage that we started real translation work.

Different groups in Blanchardstown came up with different translation tools. These and their associated DLLs were copied and shared around the organization. Anna Brady developed a DLL translation framework called Pebbles. Another in-house tool, Utah, compared old and new DLLs and leveraged previous translations. With every new build of a product from the US, the amount of translation work required got smaller and smaller.

We tested new products with sample data that included content that had been problematic in earlier projects. By the time of the final build, we knew every file that had been changed and had tested it.

One of our biggest projects compressed the timeline for releasing localized versions of new products simultaneously with the American release. At the start the gap was up to 60 days. We got that down to zero.

Symantec kept buying in more products and the software group in Blanchardstown grew rapidly. It employed about 140 people in the mid-1990s. Our building had to be expanded twice so that we could all fit in. The company now ran separate divisions for its security applications, utilities, development tools and communications software. John Kennedy took charge of localization for utilities, while Kevin Hogan managed the security products. I was responsible for everything else.

I ended up managing a whole suite of translation and testing tools. Apart from Segue Software‘s automated test technology, very few of these tools came from outside Symantec. We created a very good quality assurance process based on Lotus Notes. We did a lot of internal training and had a checklist of localization skills that we used when bringing in new people. We contracted out work to external translation services and used DiskFax to send the software masters to the US development teams for production. It evolved into a very sophisticated organization.

Every product had its own quirks. For example, pcAnywhere was challenging because it controlled a remote window on the PC. And the Hebrew version of Norton Utilities deleted files at random. It took me and my team three days to work out that the system was treating any directory with over 127 files and with Hebrew or Japanese characters as corrupt.

Earthworks was a Symantec product that only the company’s own employees could use. This manual guided them through its collection of localization software.

We asked Symantec’s tech support people in the Netherlands to test our localized products. And, whenever one of the software teams in the US was close to shipping a new release, we would send over two of our test engineers to take a look at it. That gave us massive insight and greatly improved product quality.

I also got involved with the corporate mergers and acquisitions team. They needed someone who could reverse engineer a product, inspect the strings and libraries, look for multiple forks from the original code base, work out what it would take to localize it and how much that would cost. Sometimes I would buy a product off the shelf and rip it apart. I couldn’t tell anyone what I was up to.

Something else I tried to do was to simplify the localization process. By recording every keystroke we got to know how people were using our tools and found that some of the functions were more complicated than they had to be. I brought in a traffic light system to show whether a new release was ready to go or needed attention. I also became the ‘millennium tsar’ for code testing. I told everyone what they had to test for Y2K issues.

The localization business was a small world and the centre of the universe was Lotus. SLIG provided a forum for inter-company contacts. I was not one of its founders, but I got involved with the association during its second or third year. I used it to identify people that Symantec might want to recruit. It took four or five years before SLIG started talking about standards. Certain people would never share their knowledge. Paul Quigley, though, was one person that I got to know through technical discussions. He had a tool at Oracle, called Hulk, which was the equivalent of Pebbles at Symantec.

By the end of the 1990s we had brought all of Symantec’s localization tools together under the name Earthworks – a collection with its own manual. We also produced a Symantec guide to character set testing, localization guides and a guide to doing things faster.

Damian and Paul Quigley formed Key Performance Solutions in 2002.
Photograph by Damian Scattergood.

I had been with Symantec for several years and was starting to get bored. We had cracked the localization process and I wanted to get into the internet business. I joined Worldport Communications in 2000 as director of operations, but stayed there for just one year. The company was constructing a 144,000 square foot data centre when several competitors were also racing to build data centres on the edge of Dublin. Most of them were bound to go bust. I did not enjoy my time at Worldport, so I left.

My next project was a recruitment web site for the localization community. I wanted to know who every person in the industry was and what every company was doing. The LocalizationWorks.com site still exists but the service is not active.

Then I got talking to Paul Quigley and we came up with a new business idea. We both saw opportunities for further automation of the localization process. In June 2002 we formed Key Performance Solutions to evaluate and re-engineer the processes in client companies and began to look for funding.

We went to the Localization World conference and got talking to Hanspeter Siegrist, the CEO of the STAR Group, a family of more than 30 translation services and technology providers. STAR had grown to be one of the top ten players in the localization industry. We quickly agreed that our new venture would join this group and signed a deal couple of weeks later. We launched STAR Translation Services with Paul and me as part-owners and assisted the group to fold all of its technologies into a single workflow system. I am still managing director of STAR’s organization in Dublin.

Last edit: May 2023

© Damian Scattergood 2023