Technology is inextricably linked to today’s language business. Decades ago a translator may have been able to work alone on his or her typewriter, but now vast systems are needed to manage every aspect of language deliverables.
Here are some of the most important language technologies and systems in use today:
Content Management Systems (CMS)
These systems help create, manage, and publish content, particularly in larger companies. The best products today support the creation and management of content in multiple languages. Increasingly, CMSs must connect to other technologies, such as translation memories and terminology databases.
Translation Memory (TM)
These systems store translations as human translators create them. The paired source-target translations can then be called up later for re-use when the identical or similar text appears. TMs are standard tools in localization nowadays because they vastly increase the efficiency and consistency of translation work, but they do not translate automatically on their own. They are only as good as the human translations that have been stored in them, and they only help to translate text that is the same or similar as previously translated texts. Consequently, highly original texts (such as marketing collateral) generally benefit little from TM.
Machine Translation (MT)
Machine translations are produced in a fully automated way without human intervention based on software algorithms. Because of MT’s speed and automation, it is often used to translate vast amounts of information involving millions of words that could not possibly be translated the traditional way. MT output can vary considerably in terms of quality; the best quality is obtained by MT systems that have been trained specifically for the domain and language pair required. Untrained MT systems may often produce garbled or comical results.
There are two basic types of MT systems:
Rule-Based Machine Translation (RbMT)
This type of system relies on intelligent algorithms coded into software based on grammar, syntax, and other rules.
Statistical Machine Translation (SMT)
These systems rely on pattern-matching against vast amounts of reference texts to find translations that are statistically most likely to be suitable.
Increasingly, some combination of the two above systems is now being used (these are called “hybrid MT,” or HMT systems).
Translation Management Systems (TMS)
TMS technologies are designed to streamline and accelerate the translation workflow. They range from simple portals for submitting source content to a language supplier and receiving back the translations to enterprise-wide complex systems that automate most of the handoffs between clients, LSPs, project managers, translators, editors, proofreaders, quality assurance staff, reviewers, terminologists, and more. TMS systems may or may not include built-in TM. Some systems also include vendor management tools to make it easier to qualify translators, assign work, and pay suppliers.
With a broad array of technologies now available to help with translation, each providing an elegant “island of efficiency”, the need to connect these various solutions together in seamless, end-to-end networks has become more pronounced. The more the processes can be integrated, however, the closer the “content translation lifecycle” will come to near total automation – providing tremendous scalability, reliability, and security. Connectivity platform products are now available to connect all types of disparate systems – typically a CMS to a TMS or an MT Server – automating the flow and routing of content between the various systems.
For more terms used in localization, see the section Localization Definitions.