Translate an XLIFF File into Spanish

Punctuation, capitals and length in Spanish segments

Spanish opens questions with ¿ and exclamations with ¡. In XLIFF, that opening mark has to sit correctly relative to inline tags. The source Delete <g id="1">all files</g>? becomes ¿Eliminar <g id="1">todos los archivos</g>?: the ¿ stays outside the tag at the start of the sentence, and the tag keeps its place.

A common trap is a sentence split across two trans-unit elements (or two segment elements in 2.0). The ¿ then belongs to the first, the ? to the second, and each looks incomplete on its own. That is not an error, but check it on screen.

• No title case: "Save As Draft" becomes "Guardar como borrador". Days and months are lower case (lunes, enero).
• No space before ? ! : ; even if your source is French.
• Spanish runs longer than English: "Sign in" becomes "Iniciar sesión". Narrow buttons and tabs are the first to overflow.

Placeholders, tags and Spanish word order

A single tag moves freely: <x id="1"/> files deleted can become Se eliminaron <x id="1"/> archivos. Trouble starts when two tags need to swap. The English possessive <ph id="1"/>'s <ph id="2"/> ("Marta's project") calls for the reverse order in Spanish: "el proyecto de Marta".

Manuce keeps inline tags in the same order as the source and checks every translated value against it. If no wording respects that order, the source value is kept and the report says so. If you own the source strings, avoid such constructions or split them.

• Plurals: if your sources hold ICU messages, Spanish uses the CLDR categories one, many and other; many covers large round numbers ("un millón de archivos", with de). Check that your library receives the branches it expects.
• Gender: "Welcome, <ph/>" forces a choice between Bienvenido and Bienvenida. A note in the file describing the context helps, as notes are passed to the translator.

Spain or Latin America

The same label differs by audience: ordenador or computadora, móvil or celular, fichero or archivo. Whether interface messages use informal tú or formal usted also varies by country and brand.

Because only the values change in the returned file, the target-language attribute stays exactly as you uploaded it. Set it beforehand: target-language="es-ES" on the file element in XLIFF 1.2, trgLang="es-419" on the xliff element in 2.0, or es-MX for a Mexican audience. Your import tool relies on it to file the translation.

Then read the free sample with one question in mind: do the vocabulary and register suit your users? If the sample uses ordenador and your audience is in Argentina, you see it before paying.

What Manuce returns, and what it does not do

• The same XLIFF file, 1.2 or 2.0: each source translated in its target, the state attribute updated.
• Indentation, element order, quotes, comments, line endings, encoding and BOM preserved.
• Units marked translate="no" are respected; notes are used as context.
• Inline tags (<g>, <x/>, <ph>…) stay in the same order; a string that would lose them is rejected and flagged.
• A report: strings translated, strings left as they were, and why.

Not translated: keys, empty values, URLs, email addresses, numbers, colours, date formats, identifiers. For Spanish this matters: an MM/DD/YYYY pattern hard-coded in a value stays as it is, although Spain and Latin America put the day first, so date formatting belongs in your code.

Manuce does not add plural categories missing from the source, does not change the target-language attribute and does not produce sworn or certified translations. Pricing counts the words in values only, excluding tags and keys.

Preparing the file and checking it afterwards

1. Export a complete XLIFF from your tool (Angular, Xcode, a CAT tool…), including context notes if your tool can generate them.
2. Set the target language: es, es-ES, es-419 or a country code.
3. Mark product names and technical strings that must stay unchanged with translate="no", then check them in the sample and in the delivered file.
4. On delivery, read the report first: a string kept in the source language often points to two tags that needed swapping.
5. Import and walk through the interface: ¿ ¡ correctly placed, overflowing labels, gender agreement, stray capitals carried over from English.

In practice

1. Export the XLIFF and set the target language (es-ES, es-419, es-MX…).
2. Upload the file: the quote counts words in values, excluding tags and keys.
3. Read the free sample: vocabulary, tú or usted, placement of ¿ ¡.
4. Download the translated XLIFF and its report.
5. Import it and check the labels on screen.