Terminology to suit the master plan

Discussions held in the past on the role of terminology have brought out some facts clearly: technical communication requires an agreed terminology, good technical documentation requires defined concepts or terms that are valid uniformly and without orthographic variants in all contexts. And all companies need a documented corporate language, across the boundaries of all departments, locations and languages.

Terminology work consists of arriving at agreed solutions for this and recording them in a terminology management system or “TMS”. Professional systems provide a large number of databases for capturing terminology in a systematic and structured manner. A terminology guideline defines what goes into the terminology. All these instruments and measures are known, still, terminology has been an issue for long. Why is this so?

A plethora of interfaces

From a company’s point of view one thing becomes clear: there is a growing number of areas where terminology is used, as well as interfaces for the same. On the contrary, the tool landscape is very fragmented. Terminology management still suffers from the fact that there is exactly one application behind every solution: translation using Computer Aided Translation, CAT, has its own terminology management, just as the checking software for controlled language (CLC= Controlled Language Checker). A system for machine translation, “MT” or an electronic dictionary too have their own terminology management.

Exchange processes or cross-media applications like those that are common in authoring systems are almost unknown here. Exchange formats such as the Term Base eXchange Format (TBX) exist, no doubt, but on closer examination, these are not supported adequately in most systems.

Cross-media terminology management

The term cross-media is usually associated with authoring systems and in very general terms, it stands for the communication from one source through several channels in various media [1]. A system for terminology management can also be called a cross-media-application. Like in the case of authoring systems, various media are created. But these media are not output media for users, but media formats that serve in turn as input formats for various applications.

The software architecture of a terminology management system with cross-media functions is based on a central management system. This system interacts with other applications through exchange processes. Without cross-media functionality it is almost impossible to use terminology consistently in a large number of processes. The effort involved in the synchronization of terminology data also goes up and gets multiplied with every new application or process that is added.     

Applications that need to interact with each other are all those systems, which are used for technical documentation, translation and localization. This includes translation memory systems, systems for machine translation, language checking tools, content management systems and electronic dictionaries. Besides these, even ERP or PLM systems (ERP = Enterprise Resource Planning, PLM = Product Lifecycle Management) should be taken into account. Terminology is also used and managed in other areas of the company, such as in marketing, production, finance and HR.

Fig.: Terminology management system that also works as a cross-media application

Impact on the entire company

A terminology management impacts the company as a whole when it extends beyond the specific boundaries of the translation department, the foreign language department, and the technical documentation department and is consolidated and available across all the departments, locations and applications. Its benefits become apparent when it becomes necessary to use a uniform terminology in departments whose areas of work differ. Some examples of this are:

• Labeling of machinery, display texts and manuals
• User interfaces, help text, handbooks and sales documents
• Content of web sites and marketing material
• Names of articles in the documentation and in the customs formalities

Thus, it is worthwhile to analyze the processes and plan, regardless of whether the company is still in the starting stages or already has a full-fledged terminology. A well-defined corporate language becomes all the more important as the company’s operations become more and more international. A confused or fuzzy terminology will be reflected in the foreign languages as well, because the translator will probably assume a different meaning for the second concept and look for an appropriate term in the foreign language.

Exchange processes between the various applications should be looked at as closely as the content of the entries for planning and implementing a terminology management system that aims at supporting a large number of processes.

Managing all the information

Some of the most important criteria to be met by a system are the autonomy of concepts and a variable and customizable field structure [2]. Professional systems provide data models that are partly flexible and extensible. It is advisable to take into consideration all the interfaces to applications and processes while defining an input structure and in defining the data model, so that the corresponding information can be input. In addition to this, fields that contain the necessary information for a specific process should be made mandatory.

Here are some examples of the necessary information, depending on the application:

• Terminology for CAT systems must be multilingual. The minimum requirement for additional information should be the specification of a technical field. Pure lists of words are not relevant as terminology.
• For a reference function in the Intranet or online, as well as for translators, other semantic differentiators would be helpful. This not only includes definitions, contexts and examples of usage. Media formats such as graphics, images, audio files or videos sometimes say a lot more than words.
• For checking terminology with a help of a linguistic check, the user should be able to enter in the terminology management not only the permitted concepts or terms, but also those that are banned or to be avoided. Language checking tools also have prescribed lists for details of authorization or the status of a term.
• Further, marking up of variants also helps in keeping undesirable double entries under check, based on the following:
  Orthography, such as four way tap or 4-way-tap;
  Language version or product line, such as Geldautomat (used in Germany to denote an ATM) or Bancomat (used in Switzerland to denote an ATM);
  Abbreviations, such as AF = Anfangsfolge (Audi Human Resources, Alternativfrequenz (VW, Technik), Afghanistan (ISO 3166-1), Autofous (Minolta);
  Synonyms, such as indicator light, indicators, turn signal, or direction indicator;
  Marking a preferred term, such as Kühlmittel (German equivalent term for coolant, which is not allowed for BMW) or Kühlerflüssigkeit (German equivalent term for coolant liquid, which is preferred at BMW).
• Integration with a machine translation system requires additional grammatical information (type of word, number and gender, such as Tau, der, Niederschlag/Tau, das, Seil) and sometimes also morpho-syntactic details like the valence, such as a transitive verb like ausbauen (expand, uninstall) taking on a direct object.

What are the systems capable of delivering?

No terminology management system that is delivered as part of an overall package or as a stand-alone application is pre-configured in such a way that it has all the appropriate editing templates and interfaces for supporting all the processes mentioned above. Linking a terminology management with other systems requires preparation. And this depends on the individual pre-requisites that are to be met with for the various applications, such as the data model, fields or the contents of fields.

Nevertheless a cross-media TVS does not mean utopia. Professional systems do offer cross-media functions, various applications and suites are opening up to each other more appreciably. Despite the difficulties mentioned, standards such as TBX assume an important role where exchange is concerned.

Exchange does not happen automatically

There are a number of (XML-based) formats for terminology exchange, most of which are based on XML. The Term Base eXchange has been holding its own among these. This notwithstanding, the exchange through TBX does not happen “blindly”, as is evident from the following examples:

TBX is not quite TBX: Some applications may not support the complete standard, but just sub-sets. Fixed database models go against the grain of the flexible and XML-based TBX.

Certain processes will not need all of the input information. Multimedia data are often not relevant for machine translation, hence they will need to be filtered out in these cases.

Not all concepts (terms) are meaningful for all processes: If a TVS also manages concepts that are “not to be used” through status fields, then these concepts should not land up in a system for machine translation.

The entire terminology database will not be needed all the time: It should be possible to extract sub-sets all the time, for instance, based on the technical area. Further, experience with machine translation show that the quality does not always improve if the entire terminology base is imported without checking. On the contrary, the quality improves if it is possible to restrict oneself to the core of the corporate terminology.

Coordination and testing will be necessary in any case, to ensure that the exchange processes actually do give the desired results, namely, a centralized terminology management.

Master plan essential

Though it is usually not embodied as a central function, terminology management plays an important role. Observations made so far have shown that an ad-hoc terminology management can be successful in isolated cases at the most. In any case, it is worthwhile to take a bird’s eye view of terminology work and the benefits it can bring. Where terminology work is in place already, this amounts to a stock-taking that goes beyond the boundaries of departments, applications, locations or languages. From this, a terminology master plan can be developed.

The master plan takes into account not only the existing situation, but also points out weak areas and potential. The master plan creates the pre-requisites for enabling planned terminology work, defining the priorities the building up the content and it prescribes what might be optional and what is mandatory in the supported processes.

Fundamental decision tree

One important basic component in evolving a master plan is a decision tree. The concrete pre-requisites are captured in this tree, and at the same time, a desired target state, namely, the future perspective, is defined through it. Here is an example from a decision tree:

• Is there an ERP system in existence? If yes:
• Are concepts (terms) specified in the ERP-System? – If yes:
• Should the concepts be imported into the TVS, such as DRIVER’S BRAKE VALVE or driver’s brake valve? – If yes:
• Are the concepts coupled with an article-ID? – If yes:
• Should the article-IDs be imported into the TVS?
• Are the concepts present in several languages?
• Does the ERP-System supply “only” the umbrella term or generic concepts, such as disc, or does it describe the concrete part such as the washer?

Medium and long term perspectives

An evaluation of the decision tree will throw light on the applications, exchange processes and interfaces and infrastructure that will be needed for the terminology work, and the departments or persons who will need to be involved. A master plan can be evolved only once it is clear who is involved, what applications, functions, processes, languages or even formats are affected or will be necessary. Once the system landscape is charted out, a coherent concept can be evolved for a terminology management with a perspective on the future.

The plan also serves as a basis for the company’s financial calculations: What might be worthwhile, and what steps will be implemented when? A project plan will be defined for the concrete implementation, with sub-sections and a milestone concept with the possibility of measuring the degree of target achievement.

Summary

Even if it sounds like too much work to go through the rigmaroles of analyzing, planning and implementing a system for terminology management, the days of insular solutions and multiple data maintenance in different systems are numbered. Insular solutions have no place any more in companies where a number of sub-processes are closely intermeshed.

There is no doubt that terminology work means additional software costs, time and effort and human resources. Cost savings are very difficult to measure. But like in the case of all cross-media applications, the cost to benefit ratio improves in proportion to the utilization of additional potentials. But these potentials and multiple re-use are ultimately possible only if they are governed in a centralized system.

Literature

• [1]  (2.2.2010): http://de.wikipedia.org/wiki/cross-media.
• [2]  Saft, M. (2010): Terminologiemanagement mit System. In: technische kommunikation, H. 1, S. 44–46.
• Hennig, J./Tjarks-Sobhani, M. (2008): terminology work für technische Dokumentation; Schriften zur technische Kommunikation, Band 12.