What is a translation memory?

A translation memory (TM) is a database of text segments (sentences, text fragments, etc.) in two different languages that match each other. Having access to a translation memory and a glossary speeds up the translation process and reduces your translation costs.

Why should I get a translation memory?

Bilingual text memories allow you to reuse identical and similar text segments in future translation projects. With a translation memory as a base, translation of your documents will go faster, become cheaper and text in your documentation will become more consistent.

How to create a translation memory?

The process begins with two similar files – one with the source text, the other with the corresponding target text in a different language. To create a TM, you first need to pool as many of your original files and translated equivalents as possible. Then, these 5 steps are applied in order to create your translation resources

1. SEGMENT EXTRACTION

All text segments (basically sentences) are extracted from the source and target files to create a bilingual database with original text and the corresponding translated text.

2. SEGMENT ALIGNMENT

All segments are aligned using computer solutions to pair segments in the source and target files based on their placement, content, etc. The process is highly automated and enables very fast processing, much faster than humans and with high precision.

3. HUMAN EDITING

A human review of the result is necessary, though, in which native translators browse the paired segments to ensure they really match.

4. TM CREATION

At the end of the process, redundant segments that have no matches in neither source nor target are deleted, and then the bilingual text segments are exported to the universal .tmx format or any other format you may require (e.g. Native Trados, Trados TXT, standardized Xliff format, etc.).

5. TM QA AND EXPORT

The last step involves applying computerized Quality Assurance on the final result to ensure segments are consistent, numbers, tags and symbols match and there is no text in the wrong language.

How to update a translation memory?

Even if you do have a translation memory, it may have become outdated and not include all additions, changes, etc. that your published documents have undergone. Instead of trying to update your memories and performing maintenance of them, we have a smarter and much quicker way to keep your TMs up to date. Simply send us your original and translated DTP files and idioma’s professional translators will create bilingual text files that can be copied into your existing translation memory to update it and replace outdated text segments.
The process of TM updates includes the following:
  • Updating the segments one-to-one (source-to-target) for each TM (cleaning out duplicates, forbidden terms, etc.)
  • Verifying terminology consistency including correction by human reviewers
  • Replacing your existing TMs with the updated bilingual text files to replace outdated and faulty segments.

What does a translation memory cost?

There is no universal answer to this question. Done correctly, whenever you order translation the translator should work with professional tools, creating a memory on the fly of all translated segments. This is the professional approach, and it usually should be free.
Starting out from scratch and building a memory from old documents is also possible. Viewed in the perspective of its benefits, the cost involved to create a memory is low, typically around 1 or 2 Euro cent per segment.

How to maintain a translation memory?

In the course of translating your projects, it is important that you insist on keeping your translation memory consistent. Any error you notice must also be corrected in your TM, which is usually handled by your LSP. In addition, any change you or your collaborators make in the supplied translations should also be introduced in the TM.
Typical errors include:
  • number mistakes and typos, URL mismatches, tag orders, usage of special characters and spaces or formatting based on language style sheets
  • untranslated text, which could signal missing target text
  • text consistency such as inconsistent text segments and inconsistent term consistency
  • glossary usage/non-usage (use of forbidden terms, opposite terms, etc.)

Where can I get a translation memory?

Every competent language service provider should be able to create and process a translation memory for you. If you do not have a language service provider offering this service, you can GET TRANSLATION MEMORY CREATED HERE.

How long does it take to create a translation memory?

As translation memory creation is based on alignment of two bilingual files, the speed of alignment (and thus the TM creation) depends on the input file format. For common files such as MS Office, Open Office, FrameMaker (mif), InDesign (idml), Xliff, SDXLiff, TTC, TXT, RTF, and CSV files, computers do most of the job, while a native human translator can review the aligned result at a pace of anywhere from 15,000 to 25,000 words/hour. Alignment of PDF files is a slower process with approx. 7.000 – 8.000 words/hour.

What files can I create a translation memory from?

The file format doesn't matter too much. The documents can be in Word, Excel, FrameMaker, InDesign, Acrobat PDF, or whatever other format you may have. It is important, though, to have reliable file converters that do not mess up the files or wrongly convert foreign characters.

Can I create a translation memory from DTP files?

Yes, it is possible to extract and align segments also from DTP files, such as InDesign (idml) and FrameMaker (mif).

Can I create a translation memory from PDF files?

If the text in the different languages you want to align only exists in PDF format, this is not a problem as long as the text content has not been scanned. So yes, text in editable PDF files can also be aligned, but the alignment speed is approximately half of that for DTP files.

What is the file format of a translation memory?

Depending on your needs, we can save bilingual text segments in different formats, such as TMX, native Trados, or the standardized Xliff format. This will ensure that the output is useful to you and compatible with the TM systems you use (such as Trados, Transit, memoQ or similar).

Do I need special software or tools to work with a translation memory?

Yes, if you feel you are up to working with foreign languages youself, then you need a special software capable to read and view the particular file formats, in which your translation memory is saved. There are special software viewers available as shareware, but to work effectively with your TM, you should invest in a good CAT (Computer-Assisted Translation) solution.
Such tools can be highly productive, but it puts specific requirements on the user, not the least language awareness.
The other option is to find a capable language service vendor whom you can rely on and who will offer 100% transparency into the TM process.

What do I risk with an outdated translation memory?

If you fail to keep your TM up to date, you risk issues in the translation process, such as:
1. Inconsistent resources and terminology that slow down translation, create context errors that compromise translation accuracy and cause confusion
2. Reuse of 100% identical text segments that contain errors, which are used and published again and again
3. Your buddy Jack is not a jack inserted under a car, nor a jack in a panel... A common TM issue is the existence of different translations of the same source segment, especially for short text segments. Such segments should be kept, but their actual usage checked by native speakers.
4. If selected segments are updated in local projects, they risk becoming inconsistent with the existing segments in your TMs causing even more confusion – what should be used when and where?

What is the difference between a translation memory and a glossary?

A translation memory is a databases of complete text segments, whereas a glossary is a database of words and expressions usually of specialized character. A glossary is also sometimes dubbed a term base. While a segment in a translation memory can be short and perfectly match a glossary entry, the reverse is not the case. Usually, you should avoid keeping full text segments as entries in your glossary since this would limit its usefulness.

What is a glossary in translation?

In the translation industry, a glossary is basically a list of specific terminology that helps to ensure consistency in translation. Every language, every industry and even your very own company use specific expressions that need to be addressed and used properly in order to give your translated documents the right meaning. Additionally, glossary creation together with translation memory creation help you to significantly lower your future translation cost.

Why should I get a glossary for translation?

Working with glossaries in your organization creates unified expressions and ensures everyone understand each other. The same applies to translation: to assure we use your preferred terminology in documents we translate for you, it is important to have access to glossaries. While some companies have developed internal glossaries, many still lack them.

How to create a glossary in translation?

Your language services provider can either help you to:
  • 1. Expand an existing glossary with more terms and more languages, or
  • 2. Build new glossaries from scratch based on your existing multilingual documents.
If source expressions do not exist, the process begins by extracting the typical terms and expressions from the client’s existing source files. After this, corresponding terms from matching target documents are added to create bilingual glossaries. This can be done for one language or many. To work efficiently, it is good to check whether your language service provider employs tools, which can extract the terms from e.g. PDF files. This makes work fast and efficient, while it makes the service more affordable to end clients.

How to update a glossary in translation?

If you are not sure about the current state of your glossary and whether terminology is correct, it is recommended to run a check on it to identify potential issues. This check is performed by professional linguists who will correct and update any incorrect entries, clearly indicating what has been changed and why.

How to maintain a glossary in translation?

The end goal of having translation resources of course is to ensure consistent terminology. Once your language services provider has glossaries in place, and translates documents for you, the translation process should always include mandatory glossary checking to make sure preferred terminology is used consistently, exactly the way you prefer it.

Translators and reviewers should be warned by their CAT tools whenever a term is not used or a different one is used; glossary terms should only be ignored by stating a reason and if possible also giving a condition why a given term should not be used. These reasons and conditions should be collected in the form of a report and presented to you in connection with final delivery of projects, in order to keep terminology consistent in future translation projects.

What does a glossary in translation cost?

Done professionally, a glossary should not cost anything extra if it is done on the fly and you have already indicated which terms you want to add to the glossary. If you create a glossary up front before a project, you are likely to pay the standard translation rate for creation of the same, possibly more if the terminology is specialized.

Where to get a glossary in translation?

Every competent language service provider should be able to create and process a glossary for you. If you do not have a language service provider available with this service, you can GET A GLOSSARY CREATED HERE.

Get your TM created or checked