1: This is Software Localization not text translation

Software Localization is more than just text translation, indeed the actual text translation is often just a factor in the overall project.    Taking for granted that your product has already been internationalized, translation, localized QA both functional and linguistic, and regression stages must occur to produce a high end quality product.

2: Test plans, Test plans, Test Plans

Re-use your base language test plan during the localization process with the local languages to ensure the same level of end product Quality.

3: Provide Visual reference and/or add comments

Especially where any type of visual localization is not possible, comments in software resources that define context can really help translators.   This can be difficult to appreciate especially as the meanings will seem so obvious to developers or content managers having worked on the project for the past x no of months or even years.   Remember a translator will only have a very short time span on your project and no matter what their background will not have anything like the experience on your project that you have.

Having as much context information as possible on individual strings will help translators choose the right translation from the beginning. Most translation tools will allow translators to see these comments as they translate the strings. Using visual localization technologies such as Alchemy Catalyst or SDL Pasollo will provide the optimum solution here in terms of translator context, but in the absence of such technology or indeed in applications where these technologies will not provide visual localization capability, add screenshots where possible.

4: Avoid concatenation and overuse of single strings

A combination of words in English will most likely not follow the same order in most other languages. Concatenated strings and strings that are used in multiple contexts will have grammar and gender issues.

4: Compile your TM and Glossary across your Software and UI

We have all seen software or devises where terminology differs in the help and in the software.  If this happens in the source language version, imagine the potential issues during the localization phase.

5: Utilize your TM and Glossary on both your Software and UI during localization

Many Localization managers and development teams expect substantial Translation Memory results from applying the software TM to the User help system.   In reality though the hit rare is in the extremely low single digit percentages.   Where the TM is mainly of use is in establishing a Glossary  across both.  The low Translation Memory hit rate is due to the way Translation Memory systems utilise segment breaks to provide matches.   Software UI strings are much shorter than User Help strings and thus will provide only very low fuzzy match hits which cannot be automatically applied.