Translation tools process the text but also have to carry the format of the document. That is why the files are prepared before they are imported into the tool: the tags are separated, the fields that will not be translated are locked, the structure is preserved. When this step is skipped, the problem comes to light not in translation but at delivery.
Importing into the tool
When the files are imported into the translation tool, text and formatting are separated from each other. This separation ensures that the format is preserved. The operation is carried out in a controlled manner.
Formatting tags
Formats such as bold, italic and links are carried by tags, and the tags are visible in the text. The tags are not deleted and their positions are kept. If they are damaged, the format is lost.
Locking the fields that will not be translated
Fields such as codes, numbers and brands are locked at tool level. The lock prevents them from being changed by mistake. The practice gives the team assurance.
Variable field markers
The placeholders in template files are preserved and are not translated. If they are damaged, the system output breaks. The markers are locked.
Segmentation settings
How the text will be divided into sentences is set, and it affects memory matching. Wrong segmentation lowers the match rate. The setting is checked.
Marking repeated sections
Repetitions within the document are detected with a tool and resolved once. This provides both speed and consistency. The result is reported to the team.
Hidden content
Hidden rows, columns and slides remain in the file and may need to be translated. Their presence is checked. The decision is made with the client.
Comments and tracked changes
The comments and tracked changes in the file are cleaned out or handled separately. If they get mixed in, they spoil the text. The operation is carried out at the outset.
Headers and footers
The content at the top and bottom of the page sits in separate areas and is liable to be skipped. These areas are checked in particular. It is common for them to be left out.
Footnotes and endnotes
Notes flow separately from the main text and appear separately in the tool. The numbering is preserved. The links are tracked.
Table structures
When tables are imported into the tool, the cell structure is preserved. A damaged structure may become impossible to repair at delivery. A check is made.
Embedded objects
Tables and charts embedded in the document are handled separately. These may not be imported automatically. Their presence is listed.
Font and character problems
It is checked that special characters have been carried over correctly. Corruption is noticed only later. An early test is carried out.
File naming
Working files are named according to the defined convention and match the original file. Disorder leads to files getting lost in bulk jobs. The rule is applied.
Protecting the original file
The original file is stored unchanged and work proceeds on a working copy. If the original is damaged, there is no way back. A copy is made.
Round-trip test
Whether the prepared file can return to its former format after translation is tried out in advance. If no trial is made, the problem comes up at delivery. The test takes little time.
Expectation regarding page layout
How far the layout will be preserved at delivery is taken into account in preparation. The expectation determines the method. The decision comes from the scope.
The next step
Once the files are ready, the parts of the job are distributed. Distribution without preparation produces inconsistency. The order is kept.