While each of these phases can be performed by different organizations, the record shows that centralizing the entire localization process and the supporting processes provides higher quality and faster turnaround times – simply because the linguist and technical teams within the same organization work in accordance with harmonized procedures, and the linguistic and technical phases are organically intertwined to produce the output of optimum quality.
The purpose of this centralized approach is the delivery of a language solution, a fully linguistically, culturally and technically adapted content/product/delivery ready for launch on the international markets.
Phase One: Preparation
More than 2300 years ago, Aristotle knew that “Well begun is half done”, i.e., that the preparatory stages of any project will always dictate the manner of its execution and subsequently its success or failure.
In the context of the technical phase of localization, well begun boils down to providing an editable (open) source file.
In cases where editable files exist (these generally include HTML, XML, TXT, SRT files, a wide range of documents generated in Microsoft Office suites (Excel, Word, PowerPoint, Visio, Project etc.), files generated in certain Adobe applications such as FrameMaker, InDesign, InCopy etc.), all it usually takes is to import these files into CAT (Computer-Assisted Translation) software and to get the process started.
However, in cases where editable files do not exist, they need to be created through one of the following processes:
● OCR or Optical Character Recognition
In certain cases, the materials for localization exist only in the form of scanned and/or digitally generated, uneditable documents. To make these files editable, technical localization teams perform OCR processing to generate a searchable and editable text file that will be used in the subsequent stages of the localization process. There are three types of OCR processes:
1. OCR of textual elements: performed on clean scans/digital locked files that contain text only;
2. OCR text extraction: performed on documents that include images or elements such as seals, stamps, tables, cursive or handwritten text;
3. OCR + graphical editing: performed on uneditable documents that require complex graphical processing before and/or after the localization process.
In cases where the (editable or uneditable) translated versions of documents that were not generated in a CAT tool exist, it is possible to perform OCR and subsequently the alignment process to develop translation assets (usually translation memories, but also term bases) that will enable faster, cheaper and leaner localization with improved consistency.
Regardless of the translation methods applied (machine translation powered by artificial intelligence, postediting, professional human translation etc.), the backbone of the localization process and the technology that enables the optimization of costs, speed and quality are CAT tools, and one of their most powerful features is the ability to generate translated or localized versions of documents in their native file formats, also referred to as clean files.
However, to utilize the power of this technology, we need editable files. This is why editable files are so important, and why there are dedicated teams that prepare or develop these files.
Accurate translation of any specific terms and phrases – from marketing and branding solutions such as taglines or product titles, to scientific and technical procedures and medical devices – is one of the pains in the world of localization, and one of the reasons is that frequently, the appropriate terms do not even exist in the target language(s).
The remedy is the extraction of these elements and the development of multilingual term bases before the localization process begins. This part of the process is a joint effort by technical teams, linguistic teams and the client. It has been shown time and again that this approach saves time and resources and ensures that the localized materials are consistent and of high quality.
Each project comes with specific requirements that dictate how it will be executed. In certain cases, usually on multilingual localization projects, cloud-based CAT infrastructure, which allows remote teams to share the workload, will be the best option.
However, depending on the specific features of particular CAT software relative to the project requirements, certain locally installed CAT solutions may be more appropriate.
In many cases, the choice of a CAT solution will also depend on the source file format, the file formats of any existing translation assets (for example, the existence of translation memories in the cross-platform TMX format will provide a wider palette of available CAT solutions), as well the specific quality assurance checks to be performed.
Taking all these parameters into consideration, technical teams set up the appropriate digital environment for the linguistic teams to perform their work.
Phase Two: Execution
The bulk of the execution phase refers to linguistic adaptation of the source materials, i.e., to translation and linguistic review.
Translation is by its very nature a solitary activity and it is important to understand that, as essential as they are, the translation and linguistic review phases of the localization process still only generate raw data. Translators and reviewers do not always have the context in which their translations will be used, nor it is always possible to address all the potential context-related conflicts during the translation/review phase.
Most of the time, their work is focused on the adaptation of isolated textual elements, from software strings to pure textual content of IFUs and similar documents that require further processing and that may require quality assurance or technical adaptation steps before the project is delivered.
Incorporating the translated/localized textual elements into their appropriate context (in terms of file format and layout) is the next step that must be performed.
Quality assurance (QA) steps are performed jointly by the technical and linguistic teams after linguistic adaptation.
There are three basic types of QA programs, and the type of QA program used depends on the complexity of the project.
1. Simple QA – checking the consistency, spelling and numbers in translated files in CAT software with dedicated or built-in QA tools and professional human supervision;
2. Complex QA – checking the consistency, spelling and numbers in translated files, as well as terminology and any specific requirements;
3. Full QA + DTP/Testing – checking the consistency, spelling and numbers in translated files, terminology and specific requirements + editing of clean localized files through the Desktop Publishing (DTP) process for publications or testing of localized software strings directly/via screenshots.
In addition to providing computer-assisted quality checks, the QA phase enables linguists to contextualize their translations and to edit them accordingly. One of the essential principles of the language industry is the 6 Eyes Principle, meaning that at least three different professional linguists need to check and approve the textual translated elements before delivery.
The first linguist is the translator, the second one is the reviewer, and the third linguist is the one who reviews the finalized clean file in its intended context.
DTP + Linguistic Review
Once the raw translated textual content is imported into page layout software, things may look very different. Sentence length and character width may vary depending on the target language and script, and there may be issues with text breaks, fonts and layouts.
The DTP process generates a clean translated file with these issues resolved, and such clean files require additional linguistic review that addresses any remaining issues.
In most cases, the source language for software localization is English, and English is what is known as an analytic language. This means that it has a very low percentage of words whose endings change depending on their function (for example, English verb forms do not change depending on the grammatical gender and number as in Slavic languages, among others).
This basic fact can create confusion in the linguistic adaptation phase as translators usually work on isolated strings developed in accordance with the English morphology, and one of the solutions for this issue is testing localized apps (or, alternatively, screenshots).
Software testing enables the linguistic and the technical teams to check whether this and other locale-specific aspects are appropriately addressed, such as the display of characters, fonts, dates, holidays, numbers, currencies, names and salutations.
Phase Three: Expansion
At the center of every localization process is quality. Creating an immersive user experience that will lead to the adoption of a brand/product/service is the ultimate goal of high-quality localization.
This is what we mean when we say that localization is more than translation – because it involves more than just textual elements. It is impossible to achieve this without unifying technical and linguistic (and in some cases also creative) teams under one set of values and procedures, and the result is always a complete, one-stop language solution designed to unlock the potentials of international markets.