Video games are a global phenomenon. The continued proliferation of internet access and mobile devices is expanding both the demographic of gamers and geographic penetration of games. The globalization of gaming has opened a wealth of opportunities for developers and localization teams alike. While content has become increasingly complex and new genres and form factors have emerged, the development cycle has also changed dramatically, with a significant impact on how localization is conceived and executed.

Back when titles shipped exclusively in a physical format, the development cycle ended as soon as a game went “gold.” The developer delivered the finished game in a master disc to the replication plant, which then manufactured and shipped it to game retailers, in large quantities. Usually, the finished product would be relatively bug-free with no additional content. In some cases, extra content came in expansion packs and released in a similar manner, localization’s main challenge was to produce all languages at the same time, allowing for simultaneous release (simship) of the game into a number of territories.

However, with the popularization of digital downloads, it has become relatively easy to implement changes and add extra content to games. Ultimately, this extends the overall life span and development life cycle of a game since the developer’s responsibilities no longer end on its release date. Leading game companies have figured that they need to do more to keep players engaged after launching their titles, and are gradually pivoting to a games as a service model (GaaS).

Continuous development drives continuous localization

Continuous development implies that new contents are created on a regular basis. This sets the stage for continuous localization, a relatively new process, which has made major inroads in mobile and web apps and is gradually being adopted in other industries. Continuous localization is an offshoot of agile development, whereby new content translations happen in-line with constant updates of game projects. Teams working around the clock with coordination centers in different time zones are now becoming an essential solution. The benefits of continuous localization include allowing localization teams to start the translation cycle earlier, reducing release date unpredictability and avoiding redeployment costs and risks.

In broader terms, localization includes the adaptation of all assets to the target locales: text and graphics as well as audio. High quality localized voiceover is an integral part of AAA games and is used to reproduce the immersion created by virtual environments in virtual reality games. For audio the parallel development of scripts and voice contents implies that the translation and dubbing have to move in sync with the source assets. Full access to the latest text and audio assets is therefore essential to evaluate the status of the game and plan the integration and quality assurance phases at any time.

Key tools for global development

Let’s review the main tools that developers can rely on to build a development environment able to efficiently and orderly produce and integrate multilingual content.

1. Game engine with multilingual support: The game engine is a framework used to build games efficiently. They usually include an editor to design level maps, a scripting component for the game logic and a content manager. There are many popular game engines on the market such as Unity, Unreal, Game Maker Studio and Godot aimed at different segments, from indie to AAA developers. The checklist of a good engine now includes the functions to export/import localizable assets with a click, the ability to easily encode, represent, sort and manage text in different writing systems, including Asian, Cyrillic and Arabic systems.

2. Multilingual content management system (CMS): A multilingual CMS is the centerpiece of the global development cycle. It allows easy and efficient tracking of content in multiple languages. The versatility of a multilingual CMS dedicated to games development is measured in terms of its ability to store strings with tags in any type of language, its capacity to provide context information that won’t be used in the game (side lines, pictures, metadata), the possibility to add custom tags to manage suitable workflows, the support of application programming interfaces for connection to game engines and the availability of a number of filters to seamlessly import/export data in standard formats. Such systems can be installed on premise for maximum security or in external servers and used as software as a service (SaaS) in the development ecosystem.

If well-conceived, a CMS for game development allows the developers to access the system in order to easily extract strings to be translated and reintegrated back into the engine. The same contents, if adequately tagged, can then be used by testers to flag issues on the strings, thus allowing the maintenance of dynamically changing strings in an orderly manner, with a very fast turnaround.

The role of the CMS is not to be confused with that of translation memories.

3. Translation memory (TM): A TM aids translation teams by recycling previously translated texts whenever suitable. The TM database stores both the source text and the corresponding translated text in language pairs. A TM stores words, sentences and paragraphs in unit-like segments within a database, and suggests the most suitable one to the translator. TMs then integrate with word processors, computer-assisted translation tools, and terminology management systems for optimal results. While a TM is an essential tool to recycle what was already translated in the past, thus improving consistency and speed, a TM does not maintain the history of each individual asset in the game.

The case of Asian global game development

There are a number of cases where the game uses not one but two source languages. This is typical in games developed in Asia first and then exported to the West. In this case there is one Asian source language (Japanese, Chinese or Korean) used to obtain other Asian languages and the English language, which is the source for translating into other languages like French, Russian, Brazilian Portuguese and so on.

To address this further complexity, there are CMSs (such as XLOC) that can link two asset databases in a waterfall configuration. This allows the separate but synchronized management of the two linked databases: Database 1 (DB1) manages the Asian source and the English target, and Database 2 (DB2) manages English as the source and the remaining languages as the targets. The link between DB1 and DB2 allows us to compile the latest source and target data in DB1 and trigger an English source update to DB2. The propagation action can be initiated by the project manager, who determines when the English is ready for release to the translators in DB2.

Such a sophisticated CMS ensures that all changes for both DB1 and DB2 are tracked and searchable, so that translators and managers are fully informed and act accordingly. Also, when setting up a waterfall project, string metadata are maintained and consistently carried over from DB1 to DB2.

The saving of time and gain of control are evident. The CMS becomes an essential production tool when development (and localization) iterations are repeated. Further enhancements for speed and visibility in each stage of development are:

• Combined visibility of translation progress for all languages and status tracking.

• Edit capabilities for Japanese to English.

• Edit capabilities for English to Western languages.

• Compilation of all latest source and translated data into one delivery.

Common mistakes

The request for process optimization often comes from the localization team tasked to manage the synchronized production of a number of multilingual assets created by teams distributed around the world. Common process issues include nonfinal source texts, inconsistent text and audio assets, and the churn of translation and dubbing bugs arising from builds made with outdated assets.

As a result, the localization team often selects a TM to provide a common collaboration platform to serve the translators, while the need is actually for a CMS that serves as a content repository for all stakeholders.

Other times there is the expectation that the CMS is the environment where translation happens, acting as a TM. This does not work either, due to lack of the complex translation-specific features like sophisticated glossary management, and ergonomic translation surfaces, collaboration environments allowing hundreds of translators to quickly interact with the TM.

It is safe to say that given the actual state of the industry, all three tools above are essential for the modern development life cycle of global games, and communication between them is one of the areas where optimizations could be achieved.

Global game production is one of the most exciting activities in the digital world. Never static, it has changed every year over the past four decades. Localization has become integral in the production process, and this is especially true when continuously deploying games to a global fan base, whether via PC, console or mobile.

Allowing creative choices to be made at a late stage in production and being able to iterate the localization process on a daily basis require a good technological and organizational setup. The combination of game engine, CMS and TM can help to propagate new content across a number of locales with minimal effort and maximum focus on both contents and cultural relevance. As the industry develops, a deeper integration of all processes will continue to be required, bringing the different stakeholders together in an efficient and coordinated effort.