How to keep the layout when translating a document

Why layouts break in translation

Translated text is rarely the same length as the original. The W3C's internationalization team (the group that writes web standards) gives examples: the English word "views" came out about three times as long in Italian and about 2.8 times as long in German, but shorter in Korean. Citing IBM's guidelines for European languages, it notes that short strings grow the most: text of up to 10 characters can expand by 200 to 300 percent, while passages over 70 characters grow to about 130 percent of the original.

That is why the parts that break first are the short ones: table headers, form labels, buttons in screenshots, chart captions and headings. Paragraphs usually stretch less, but enough to push a table onto the next page or a signature block off the bottom.

Other causes:

• Long compound words. W3C warns that words in languages such as German, Finnish and Dutch may not wrap in narrow spaces.

• Taller scripts. Thai, Arabic and Chinese can need more line spacing than English.

• Shorter translations. Text can also shrink, leaving awkward white space.

• Direction. W3C notes that right-to-left scripts such as Arabic and Hebrew affect the organization and direction of the whole page.

Before translation: build a flexible file

W3C's core advice is to design layouts that can stretch: let text reflow, and avoid small fixed-width containers. In a Word document, that means:

• Use styles for headings, body text and lists. Microsoft explains that changing a style updates all text that uses it. If German headings run long, you shrink one style instead of every heading.

• Use real tables, with rows that can grow. A table built from tabs and spaces falls apart when the text changes length. Avoid fixed row heights that cut text off.

• Avoid text boxes and fixed frames for body text. A box sized for English is the classic place for a translation to get clipped.

• Don't abbreviate to make things fit. W3C points out that other languages may not abbreviate the same way. If you must, give the translator a list of what the abbreviations stand for.

• Leave room. On forms, slides and labels, leave space for longer text rather than filling every line.

• Keep text out of images. Words inside a picture can't be translated in place. Supply the editable chart or diagram, or plan to recreate the image.

Tables and forms

Tables cause the most layout trouble. A few habits help:

• Set column widths that can adjust, and allow text to wrap within cells.

• Keep one piece of information per cell. Don't stack two labels in one cell with line breaks.

• Mark header rows as headers so they repeat on each page if the table grows.

• On fillable forms, check that translated labels still sit next to the right box.

Fonts and scripts

Ask the translator which fonts suit the target language before work starts. A font that works for English may not cover Cyrillic, Greek, Vietnamese accents or Asian scripts, and a last-minute font swap changes line lengths everywhere. Agree on fonts at the start.

Right-to-left languages

Arabic, Hebrew and other right-to-left languages need more than reversed sentences. Microsoft says Word's Right-to-left paragraph button, on the Home tab, appears only once a right-to-left language is enabled in Office, so the person doing the layout needs that set up. Decide with the translator how numbered lists, tables, page numbers and images should run in the translated version, and check them in the final file.

After translation: the layout check

Even a well-prepared file needs a final pass in the target language. ATA (the American Translators Association) advises checking a translation in its target format before release, because problems missed earlier show up there. It also suggests changing the font of the whole text during review to expose unusual paragraph breaks.

Go through the translated document page by page:

1. Overflowing text: cut-off cells, labels or captions, and text hidden behind images.

2. Page breaks: headings stranded at the bottom of a page, tables split badly, signature blocks separated from their text.

3. Page count: compare it with the original, especially if the page numbers are referenced elsewhere.

4. Tables of contents and cross-references: update them so the page numbers match.

5. Headers, footers and running titles: often forgotten in translation.

6. Images with text: translated, replaced or captioned as agreed.

7. Numbers, dates and units: formatted for the target audience as agreed.

Who fixes the layout?

Fitting translated text back into the design is often called desktop publishing, or DTP. ATA recommends agreeing ahead of time on whether it's charged and who does it. Settle three questions before work starts: how closely the result must match the original, who does the final layout, and in which program the final file will be checked.

Starting from a PDF

If your source is a PDF, especially a scan, the layout problems start before translation. Converters often turn columns into one stream, tables into loose lines and headings into plain text. Why PDF to Word conversions mess up formatting explains what goes wrong and how to repair it. A cleaner starting file means a smaller layout job at the end.

Start from a file built to stretch

When it launches, EasyToDecode will convert a PDF or scan into a .docx designed for translation: headings and lists as Word styles, tables as real Word tables, text boxes kept to a minimum and sentences kept whole, with the original layout kept as closely as possible. It doesn't translate the text, and very complex designs may still need a final check. See how it works or join the waitlist to hear when it opens.