Computers have been aiding humans and increasing our productivity for the last 100 years; the field of technical writing is no exception. The Linguistic Society of America states there are over 6,900 distinct languages, meaning translating written work by hand can be a tedious process. Over the last few years, computer software has been able to translate text from one language to another. This field known as machine translation is a very powerful tool in the field of technical writing and language as a whole.
Machine learning, like any tool, has its positives and negatives. First off, it’s fast, as in near instantaneously fast. For example, this blog post could be translated via Google Translate from English to French in a tenth of a second. Also, it’s cheap! Google Translate is free while more enterprise level services are still considerably cheaper than hiring a translator. Yet, professional translators around the world still have employment. This is because machine learning still has a long way to go before it can replace a human translator. Machine translators are notoriously bad at translating idioms and other language specific expressions. An example is the term “once in a blue moon” translated into any other language besides English would make little sense, the software must be smart enough to realize the author means “very rarely” and translate from there.
While I’m sure software developers are hard at work to fix these problems, there are still a few ways users can get the most out of machine translators. First is to use them as a tool to aid professional translators. A machine translator can be used first and then a translator could make any changes the software incorrectly translated. This process is in use today and greatly increases the effacing and reduces the cost of translation services. The second method is to implement a few best practices in your writing so that the software will have an easier time translating a piece of writing. These include:
- Writing simple, short sentences.
- Jargon, idioms, and rhetorical flourishes are not your friend.
- Using consistent terminology.
- Structuring content consistently.
- Avoiding culture-specific references.
Depending on the need to translate work into a variety of languages, these best practices can be considered if a translation services is not a viable option. Regardless, it shouldn’t be long before machine translation is capable of understanding idioms and human syntax in a variety of languages making translating documents easier than ever before.