Improving the Efficiency of Translating Files Containing Programming Text

Files containing programming text or ‘Do not translate’ sections are becoming increasingly common as more businesses look to localise their online projects and web pages. These localisation projects present translation providers with the challenging task of separating the content for translation without impacting the programming code. In this blog, Mission Translate reviews these challenges and their impact on our clients’ work. We then present the advantages of our latest localisation tool, MTCodeCheck, which offers an automated quality control check to improve accuracy and efficiency within the process.

The increased demand for localising files for online use

Worldwide digital communication was increasing before the pandemic as businesses saw the opportunity of engaging with larger markets spread across the globe, and Covid-19 has only accelerated this online activity. Data from McKinsey Digital reported that in the summer of 2020, we leapt five years forward in consumer and business digital adoption in around eight weeks. With lockdowns extending across the world, a large majority of businesses continue to rely on online engagement, transactions and sales to keep going, and localising this digital content for global audiences is rising concurrently.

Working with numerous leading market research agencies, where current restrictions limit face-to-face interviews and focus groups, the prevalence of requests for online quantitative studies, which contain overlays or programming text, has visibly increased.

The demand has shifted in other sectors too. For example, businesses with eCommerce sites seek localisation of their user journeys to offer their global audiences access to their products. Meanwhile, sectors such as finance, legal and healthcare assess global data from their online user journeys to deliver their services worldwide.

Localising these files present translation providers with unique challenges, quite unlike other projects.

The challenges of localising overlays

Extracting the translatable text

When files containing programming text are received for translation, the first step is marking up this text with ‘Do not translate tags’. These mark-up tags must remain untouched by the linguist team to avoid altering the programming language and causing an error once uploaded.

These tags allow translation management tools, such as our MTHub platform, to identify the sections of text that do not require translation, passing only the relevant ones onto the linguists.

Repopulating the file with the translated sections

Once the linguist team has translated the relevant sections of the file and these are approved, the translated text needs to be integrated back into the original file using translation management software.

Reviewing the file

Reviewing the file for translation accuracy at this stage requires linguists who specialise in working with overlays. Nonetheless, verifying the accuracy of the programming text is extremely challenging for several reasons:

• Small characters and spacing must exactly match the original.
• Spacing around the programming text must also precisely match that of the source file.
• It is common for files to contain small translatable terms embedded within programming text. These sections cannot be turned into ‘mark-up tags’. Therefore, they are translated manually without using a TM tool, thus increasing the chance of a minor error.

Accidental modifications to programming text can be very difficult for the human eye to identify. Checking a large file can sometimes take days!

The impact of errors

A simple missed space or incorrectly used hyphen can cause a file not to upload correctly or create errors during the user experience.

For market researchers, this could mean incomplete or invalid data collection. Meanwhile, on eCommerce sites, an error message could mean an abandoned basket and a lost sale.

The time required to rectify the error can also be costly for the brand, as they experience downtime and cannot meet their required deadlines.

A customised solution

Across all our translation projects, our linguists use CAT tools integrated into our project management portal, MTHub. These tools provide us with our first stage quality assurance check, suggesting potential errors to the linguists that they can choose to accept and rectify as appropriate.

However, with the intricacies involved in this translation process, we recognised the need for a second line of defence; technology to improve the accuracy and efficiency of the overlay checking process. This requirement led Mission Translate to develop our custom-built quality assurance tool – MTCodeCheck.

This tool provides us with an automated review of the project, once exported. It meticulously completes a full-check of the file to ensure the programming text is identical to the source file in every cell, including reviewing the small sections that could not be marked up with ‘Do not translate’ tags, which were handled manually.

Via a check that takes minutes instead of hours, MTCodeCheck provides us with a report identifying the possible errors, their location, and suggested correction.

The checks performed by MTCodeCheck can be customised according to the client’s requirements and the type of file reviewed, and the process can be run multiple times until a ‘clean’ report is delivered.

A significant time-saver, MTCodeCheck also provides clients with the reassurance that their file will work correctly upon receipt and users of their web pages will enjoy an error-free experience.

Using MTCodeCheck

MTCodeCheck is used on all files containing programming text that are handled by Mission Translate, meaning we can pass these time- and cost-benefits onto our clients.

Clients can also procure the use of MTCodeCheck on files that are already localised and require quality control checking before use.

For further information on how to access this service, please get in contact with our team.