The History of Trados: Commercial CAT Tools from ALPS to Today

The history of Trados is not the story of a single company; it is the story of how the translation business became industrialised. Forty years ago, translation software was a helper program written by a few engineers for a narrow group of customers. Today the same family of software is the standard infrastructure of the global translation supply chain.

Contents

This article follows the road that begins at ALPS and runs to today’s cloud-based platforms. The aim is not nostalgia: knowing which need the tool you use was born out of also shows you what to expect from it and what not to.

At the end of the story there is a concrete outcome: how your quotation is calculated. Forty years of technical accumulation shows up on your invoice today under the name “repetition discount”.

The first generation of commercial CAT tools: ALPS

When the ALPAC Report of 1966 largely cut the public funding set aside for the goal of fully automatic translation, researchers turned to a more modest objective: making the translator’s work at the screen easier. The first commercial products of this turn appeared in the United States in the early 1980s.

Automated Language Processing Systems (ALPS) is the most frequently cited name of that generation. In the systems the company offered, the translator saw the text on screen while the software provided dictionary lookup, term suggestions and similar sentences that had been translated earlier. The decision always stayed with the human.

This model set the pattern for every CAT tool that followed. The software’s job was not to translate but to remember and to remind. ALPS never reached the commercial scale expected of it and in time turned into a translation services company; but the idea of the workflow it left behind remained.

The history of Trados: a service company in Stuttgart

Trados was founded in Stuttgart in the mid-1980s and, in its early years, was not a software house but a language services company. It began developing software in order to raise the productivity of its own translators. This detail matters: the tool is the product of a daily bottleneck, not of a theoretical design.

The bottleneck was this: large corporate customers had every new version of their user manuals translated, and the difference between versions was usually a small part of the text. Even so, the whole document was translated again, and invoiced again, every time.

The solution was to write every translated sentence into a database together with its source and its target. When a new version arrived, the software split the document into sentences, searched the database and suggested whatever it found. The idea looks simple, but making it work required the text to be broken into parts reliably and similarity to be measurable as a number.

MultiTerm and Translator’s Workbench: two products, one logic

The company’s first separate product was a terminology management program: MultiTerm came onto the market around 1990. It was followed in the early 1990s by the translation memory engine Translator’s Workbench. Together the two made up the two basic components of what we now call a “CAT tool”.

The distinction still holds today: a translation memory stores sentences, a termbase stores words. The first delivers speed, the second consistency. When an organisation’s brand name, product code or legal definition stays the same throughout a document, there is usually a termbase behind it.

Workbench’s critical design choice was that it ran inside the word processor. The translator stayed in the environment they were used to while the software generated suggestions in the background. This choice accelerated adoption, but it also created a dependency on the desktop operating system. The trace of that dependency persists today: most desktop CAT tools run on Windows, and Windows holds a 78.43% share of the desktop operating system market (StatCounter, July 2019).

Competitors take the stage

From the mid-1990s onwards the market ceased to have a single player. Déjà Vu was developed in Spain, Wordfast in France, and memoQ in Hungary in the mid-2000s. Each of them targeted a different gap.

Wordfast’s starting point was price and simplicity: a lightweight add-on running inside the word processor. memoQ, for its part, put teamwork and file exchange between tools at the centre, and adopted features such as live memory sharing early on. Déjà Vu drew attention for its subsegment matching.

This competition worked out well for users but created a cost as well: every tool brought its own file format. Common formats (TMX, XLIFF) were developed for exactly this reason. Even so the problem did not go away; in the case study, 90% of translators reported that they had experienced compatibility problems, and 70% of those reported that the problem recurred frequently (Pietrzak & Kornacki 2021).

The second effect of competition was felt on prices. Licensing shifted from a one-off purchase to an annual subscription, and free or low-cost options multiplied at entry level. Even so, the price of the tool continues to rank second, at around 40%, among the reasons given by translators who do not use CAT (Pietrzak & Kornacki 2021, Table 6.11). Cost remains one of the real obstacles to the technology spreading more widely.

The age of mergers and the market today

From the mid-2000s onwards the sector consolidated. Trados joined SDL in 2005, and the combined product continued under the name SDL Trados. In 2020 SDL was acquired by RWS, and the product family is developed under that roof today.

Consolidation cuts both ways for the buyer. On the positive side, a common working language forms around the dominant tools and changing supplier becomes easier. On the negative side, the number of players that set licence prices and file formats falls.

Tool	Usage share (2021 case study)	Standout feature
SDL Trados	43.3%	Widest file format support, prevalence in corporate use
memoQ	29.1%	Design focused on teamwork and compatibility between tools
Wordfast	10.6%	Low cost, lightweight installation
Other / cloud tools	The remaining share	Browser-based working with no installation required

The figures in the table come from a case study carried out with 141 Polish translators (Pietrzak & Kornacki 2021, Table 6.2); they should be read as a strong indication rather than as a precise market share worldwide.

The cloud generation: from the desktop to the browser

From the 2010s onwards a new generation took the stage: translation environments that run in the browser and need no installation. These tools targeted two chronic complaints of the desktop generation – the mess of installation and the need to move files between teams.

In the cloud model the memory, the termbase and the project files sit on a server. Translator, reviser and project manager connect to the same resource at the same time. Version confusion largely disappears and the time it takes to get started falls. The requirement for a Windows machine with everything installed also loosens, which is an important opening for translators working on Mac and Linux.

The gain is not free. The document no longer sits on the client’s or the agency’s computer but on a service provider’s server. Which country the server is in, how long the data is kept and whether the machine translation connection is switched on are decisive questions where confidentiality is concerned. As in the memoQ example cited in the book, the tools themselves have begun to warn users about data protection obligations (Pietrzak & Kornacki 2021, p. 25).

Today desktop and cloud live side by side in the industry. Desktop tools are still strong for long-lived documents with heavy formatting; the cloud comes to the fore on fast-moving projects involving many people. The choice should be determined by document type, not by ideology.

The Türkiye context

The market in Türkiye has witnessed only the last third of these forty years of development. CAT tools entered the country in the 2000s, mostly through the subcontracting requirements of agencies abroad. In other words, the tool arrived not out of a search for productivity but out of a client requirement.

This has had two local consequences. First, knowledge of the tools was for a long time seen as “the condition for winning work from abroad”, and the habit of building memories for domestic clients took hold late. Second, because licence costs are tied to the exchange rate, they became a heavy item for freelance translators. The case study shows that 91% of translators paid for their licence out of their own pocket; in Türkiye the weight of that burden is even more pronounced because of the exchange rate.

Against that, the families of document most often translated in Türkiye – tender specifications, ISO documents, pharmaceutical licensing dossiers, bank contracts – have a high repetition rate. Organisations that build a memory infrastructure make substantial savings on these documents; those that do not go on buying the same text again for years.

What does this mean for you?

The buyer-facing side of this history is this: which tool your supplier uses is your concern, because the choice of tool determines both your invoice and your freedom to move to another supplier one day.

1. When you request a quotation, ask which CAT tool is used, and get the answer in writing.
2. Have it written into the contract that the translation memory will be delivered to you in TMX format at the end of the project.
3. Ask for the term list separately as well; together these two files form a portable asset.
4. On second and subsequent jobs, insist that the repetition analysis be shared with you as a report.
5. If you are offered a workflow locked to a single tool, ask about the cost of leaving it at the outset.

These points do not take the place of trust in your supplier; they make it measurable. To see what kind of memory set-up makes sense for which document types, take a look at our services page, and you can request a quotation for your own stack of documents.

Frequently asked questions

When and where was Trados founded?

Trados was founded in the mid-1980s in the German city of Stuttgart as a language services company. It moved into software development in order to raise the productivity of the translation service it provided to its clients; its products were put on sale separately in the early 1990s.

Why is ALPS considered an important beginning?

In the early 1980s ALPS developed the first commercial systems in which the translator worked interactively at the screen. Built on a division of labour between human and machine rather than on full automation, this approach became the basic model for every CAT tool that followed.

Which CAT tool is most used today?

In the case study carried out with 141 translators, SDL Trados came first with 43.3%, memoQ second with 29.1% and Wordfast third with 10.6% (Pietrzak and Kornacki 2021, Table 6.2). The market is concentrated but not dependent on a single tool.

Can different tools open one another’s files?

Partly. Common formats such as TMX and XLIFF were developed for this purpose, but losses of formatting and tags are frequent in the transfer. In the case study, 90% of translators reported that they had experienced compatibility problems, and 70% of those reported that the problem recurred frequently.

Sources

• Pietrzak, P. & Kornacki, M. (2021). Using CAT Tools in Freelance Translation. Routledge Focus, Chapter 6, Table 6.2.
• Garcia, I. (2015). “Computer-aided translation systems”. Chan Sin-wai (ed.), The Routledge Encyclopedia of Translation Technology. Routledge.
• StatCounter Global Stats (July 2019). Desktop operating system market shares.