I would like to create a new glossary. What is the best way to start?
Glossaries are a great way to customize your translations and to define the translation of specific terms, such as product names. There are three ways to create a new glossary:
In this article, you'll learn more about DeepL's recommended best practices for the glossary generator and manual glossary creation. For more information about how to upload glossaries, visit this article.
Using the glossary generator
The easiest way to create a new glossary from scratch is to upload your already translated language data to DeepL's glossary generator. To get the most out of the glossary generator, you can follow these best practices:
File types
Files in .tmx or .tsv format currently provide the best results, as these formats provide the clearest indication of which source sentences correspond to which target sentences.
File structure
Files in .tmx, .tsv, or .csv format should only contain phrases and sentences in the languages of the pair you’ve selected.
When uploading .csv and .tsv files, make sure to add "Source" and "Target" (separated by either a comma or a tab) in the first line of your file. Semicolons are currently not supported as separation character.
Example files:
- .csv
- .tsv
- .tmx
File size
For .tmx or .tsv files, there should be a minimum of 1000 sentence pairs (source and target sentences). For example, if creating a glossary for English and German, there should be 1000 sentences in English and 1000 sentences in German.
For .docx and .pdf files, there should be a 1000 sentences each in a file.
Further best practices
- Glossary pairs should be used consistently in the file
- Obvious or generic translations are discarded, as it is not a dictionary-like translation.
- The time for the glossary generator to produce entries will depend on a variety of factors, like size of the file
- The glossary generator should yield an average of 20 entries per 1000 sentence pairs, but the exact number will depend on the domain-specificity and originality of files
- Trying the glossary generator in both language directions may give additional entries (i.e. from English to German and German to English)
Manually creating glossaries
Alternatively to using the glossary generator, you can also add glossary entries manually. As you collect terms for these entries, keep the following in mind:
- Brand-specific terms, such as company jargon or product and service names
- Industry-specific terms
- Audience-specific terms
- Terms that should not be translated
If possible, you can also tap into your company-internal resources. Reach out to your technical, sales and marketing teams to enrich the collection of terms you include, so you have a good base of terms for creating your own personalized glossary.