With the emergence and growth of instant translation on the web, language service providers are constantly asked if they are worried about the future of their industry.
Despite all of the benefits of machine translation (MT), this burgeoning technology is still far from perfect. MT only works for certain language pairs that have sufficient data, and in addition, has a hard time dealing with the complexities of languages.
MT can render sentences completely unintelligible when translated into a language that employs different sentence structures and grammar rules. Sometimes, the translated document may end up looking like a string of random words, which can only be deciphered by a human being who knows both the source and target languages.
In other words, when accuracy and quality matter, a human translator is very much needed when MT is involved.
Quite often for MT-related projects, linguists are used to clean up the translated text, although the demand for exclusively human translation still far outweighs that of MT. Industry wide, people recognize that although MT is the best solution for a specific, small segment of translation projects, MT is still not a desirable alternative to translation by an educated, well-trained and quality-controlled human translation team.
Use of Machine Translation in the industry
Contrary to popular belief that the translation industry feels threatened by machine translation, most every translation company does offer machine translation to their clients. Certain characteristics are normally required for a document to be considered for machine translation:
- Large amount of text
- Short turnaround time
- Informal audience
- Not for publication (research purposes)
- Low expectation of resulting quality and accuracy
- Relatively small budget for a large amount of content
If a project meets all of the above criteria, machine translation services may be a viable solution, and the rates and turnaround times can be a fraction of the cost of human translation. (Gigabytes of data can be completed in days, instead of weeks or months)
MT is less expensive than human translation because the results are of much lower quality. While human errors are not an issue with machine translation, style and grammar are. However, if one simply wants the gist of a document (e.g. find keywords, names), machine translation is an acceptable option.
Few machine translation databases are as advanced as Google’s highly used translation tool, which at this point, includes 64 languages. The accurate conversion from one language to another depends on the amount of data one has to build a database of words, terms and phrases.
Use of Translation Memory in the industry
Human Translators commonly use Computer-Assisted Translation (CAT) tools on a daily basis to help complete translation projects. CAT tools, of which Translation Memory is the most common, help translators build their own database of terminology.
Translation memory is a commonly misunderstood tool that has risen in popularity in the past decade, not only with translators but with translation clients. Clients often request the use of TMs to ensure consistency of terminology across multiple projects, and the use of these tools is encouraged by translation companies.
TMs are a way for translators to record past translations for future use. While this was done manually in the past in the form of glossaries and tables, programs now exist to simplify the process.
Programs such as Wordfast, MemoQ, Trados and SDLX have made it simpler to use consistent terminology through long texts and multiple translations. These CAT tools save translated words, phrases and segments into data banks, where they are accessed when the same terminology shows up in the future. A current trend is emerging, where teams of translators can access and update the same TM database simultaneously, which can be a tremendous advantage for large projects with short time frames.
As a result, translators can avoid doing hours of research on a project they worked on days or even years before!
Between TM, MT and CAT, the translation industry is swimming in acronyms. Understanding the benefits of these tools will not only keep translation companies competitive, but it will help them, in this technological world, to stay afloat.