As online marketing has become increasingly important to all businesses, translation services too have evolved and expanded.  One of the major arguments today, is whether machine translations are good enough now, or indeed, will ever be good enough to provide a suitable standard of service.

Recently, it has been revealed that Google has begun charging businesses to receive machine translations.  Throughout this article, we will look into the evolution of translation, and indeed machine translation, in a bid to discover whether such a cost can be justified.

How Have Translation Services Evolved?

One of the major ways that professional translation services have evolved is through localization.  This is both within a geographical and an industry-specific sense.  This means that a translation service will be conducted with consideration to both the linguistics and culture of a specific location.

Taking the example of French, which is an official language in 29 countries, it is quite clear that the French spoken in West Africa is different to that spoken in Europe, Canada, South-East Asia, the Caribbean and all other French-speaking regions.  Indeed, the dialects and cultures can even vary hugely even within a single state.

With regards to industry, many professional translation agencies now ensure that they have experts within specific industries, such as; medicine, engineering, IT, Law and Science, where specific vocabulary and documentation is used.  This is especially important within sectors where health and safety issues can be the result of poor translation services.

With this in mind, the very best translation agencies now have a comprehensive network of translators, spreading throughout a variety of sectors and locations.  Many also have systems in place to ensure that each translation job is given to the most capable team of translators.  This is especially important when dealing with niche texts, minority languages or rare language pairings.

Machine Translations

The origins of machine translations, hails back to the 17^th century and Renee Descartes, although the first demonstration is thought to have been conducted by a Georgetown research team, in 1954.  As with many technologies, its real growth came alongside the rise of the internet, and in 1997, Alta Babel-Fish was racking up half a million requests per day.

Like many other aspects of the internet world, it was eventually Google who took over with regards to online machine translation services, with the launch of Google Translate, which, as of 2012, claimed to translate enough text each day to fill around 1 million books.

The argument, which rages on, has always been with regards to the quality and accuracy of such translations.  Indeed, a simple experiment of translating a phrase through several languages and then back to the source tongue can prove its inadequacies.  A comedy group even recently put the ‘Fresh Prince of Belair’ theme tune through the Google Translate ‘filter’ and then performed the resultant songs.

Businesses Paying For Machine Translations     

In February 2012, it was announced that Google Translate would be releasing a paid version of the software, for business and commercial use.   Jeff Chin, the manager of Google Translate, announced that companies would pay $20 per million characters – working out at approximately $0.05 per page.

Although this is, of course, much less than one would have to pay for a professional translation service, the quality too is far removed.  Indeed, simply by searching for Google Translate on Google, you’ll find that the majority of the search results are related to humorous anecdotes, which highlight the software’s shortcomings.

Quite simply, when it comes to professional translation services, you can’t beat a human expert.  Even though, of course, human error is a possibility, it is much less likely than machine error, where there is a clear lack of cultural, linguistic and specialist knowledge, and with professional translation agencies evolving their services, for any business serious about its online marketing, machine translations are, quite simply,  a no-go zone.

Mobile telecommunications is spreading to every corner of the globe at an incredible rate.  For example, in 1998, there were fewer than four million mobiles on the African continent, whilst today, there are more than 500 million, or half the entire population.  With smartphones becoming commonplace in all major cities, there is also now a huge demand for apps.  Africa alone has approximately 2000 languages, with Nigeria boasting an incredible 528, and thus even when focussing on a relatively specific area, it is imperative to consider language and accessibility.

This presents app developers with a new challenge, a challenge of not only overcoming language barriers, but also cultural differences and sensitivities.  It is not only Africa, of course, smart phone technology is becoming more and more influential in both developed and developing countries around the world.  Android devices are now also experiencing massive growth, being deployed at a rate of about 1 million units per day, thus further expanding the potential market for apps.  It is imperative for new apps to consider spreading their net through language, and existing apps also need to become multi-lingual if they are to truly maximise their sales potential.

If you look at the current top twenty-five apps, you will see that fifteen of them are games.  The rest are spread out between sports, music, communications, photography, and navigation.  These are all universal in appeal, and their success is partly as a result of their accessibility.  In fourteenth place on today’s list, you will find ‘Angry Birds’, one of the most frequently downloaded Apps in history.

‘Angry Birds’ is currently available in seven languages, and if this was not the case, it would never have seen the success that it has.  It has now been satirised in a viral video of the failed Palestinian-Israeli peace talks, used in Malaysia to promote an attempt to set a world record for the largest number of people playing a single mobile game, and featured in adverts in Finland, Japan, The US, and Spain among others.  This highlights the universal appeal of the game, and how it has benefited from being accessible in a number of different languages and countries.

With independent programmers flooding the market, the success of Apps is hugely reliant on their ability to appeal globally.  Indeed many emerging markets, have limited programmers, and thus there is a huge gap in the market for apps in their languages.  App entrepreneurs have two options available to them.  They can either target specific emerging markets where few apps currently exist, or indeed simply use translation agency services to ensure that any app they design has a universal appeal and accessibility.

According to a recent report by Distimo, Asia now boasts the second-largest global app market, in China, and also has many other countries whose app download rate is growing significantly faster than more established markets.  To draw comparisons with another industry growing in Asia, the film industry now focusses most of its funding on films that will have an appeal in the Asian market.  There is, for example, limited funding for comedies, which are very culturally specific.  The App industry needs to follow suit, and consider the global appeal of an app before beginning development new projects.

In brief, as with many businesses in the world today, to be truly successful in the App industry, global appeal is crucial.  It is absolutely essential to study recent statistics and developing trends, and remain one step ahead of the game.  Whether it is with new apps, or existing apps, translation services should become a central part of what you do. If you are truly to fulfil your apps potential, it is no longer enough to be available in one or two languages, rather you should be looking to spread the net as wide as possible.

Any person keen on a career in translation services should be aware that proficiency in different languages is not the only skill needed. Familiarity and expertise in using CAT tools is also important to add that extra edge to your knowledgebase. CAT tools can expedite the time taken for translation. Besides they also provide a consistency in the translation services, thereby ensuring that the quality standards are met.

For a beginner in the field of translation services, selecting the right CAT tool can be a difficult choice. There are many CAT tools available currently, but before you choose any one of them you need to make sure whether they serve your translational needs. Some of the freely available CAT tools can act as a good starting point since they offer good scope to try, to experiment and to learn the tool without making any investment. Some of these are Wordfast Anywhere, OmegaT and Anaphraseus. One can also evaluate the suitability by downloading the trial versions of CAT tools such as Trados, Alchemy, MemoQ, or Fluency. Across is a tool that works for all Microsoft applications.

Another factor that needs to be considered before you decide on a given CAT tool is whether the translation services you routinely offer are related to heavily formatted word documents. Also if you are working on ppt files needs to be checked. Some CAT tools do not support the language pair that you might be required to work on, so check this fact beforehand. Also whether support is there for languages with a non-Latin script or right-to-left writing style is important.

Using CAT tools to support translation services is an effective way of leveraging technology to simplify your routine tasks. The more you train and learn the usage of these tools, the more value will be derived. However, do remember that CAT tools use the basic principle of translation memory, so it might take up a bit of time to build up your own database, and you might need to collaborate with other translators.

Translation memory has become an important tool frequently used by the best translation companies in order to provide quality and consistency in its services. Basically, translation memory builds a database in which it stores all previously translated text. The original text and its corresponding translation constitute a pair. These pairs can be used during manual translation to speed up the translation time and also to ensure that all translations are consistent with previously translated documents.

Using artificial intelligence based logic to reduce manual effort is not something totally new. This principle has been incorporated in translation memory software. Thus all earlier knowledge and work is saved in a repository and suggestions are offered to the translator based on the translation unit pair retrieved from the database. It is up to the translator to accept or reject the suggestion.

We see that translation memory tools can effectively complement the process of human translation. The time and effort spent in creating a high quality piece of translated text is available for future within the translation company when translation memory technology is used. Translation memory software helps the translator cut down on repetitive and tedious tasks and to rather focus on translating the right context and semantics of the work. It is especially useful in situations where the product is undergoing a quick revision and the associated documentation has to be translated quickly.

A translation company experienced in the area of linguistic services will leverage the power of translation memory tools. However, it has been reported that some translators prefer no to  the use these tools. On reason is that some translation companies insist on paying the translator based on the amount of text translated, and not paying or discounting the text that has been converted through translation memory software. This approach is not effective in actual practice, because the translator spends equal amount of time or maybe even more,as they evaluate using the suggestions offered by the software.

Any translation agency attempting to market its services will tend to highlight its services be in terms of quality or speed. While it may be tempting and simple to compare the speed of translation services since it involves only numbers to evaluate, interpretation and evaluation of quality is rather subjective. Translation software is used to assist the translator and improve the speed of matter he or she can translate within the shortest possible time.

Translation software can offer advantages as well as disadvantages for a translation agency. One needs to have a fair idea of both the pros and the cons to be able to derive the maximum value from translation software.

Some of the important advantages are that translation software helps to reduce repetitive and mundane translation tasks, thus enhancing the translational speed. A database of the translated terminology can be used as an online dictionary. In fact this principle is extensively made use of in the translation memory software. Fuzzy logic based principles help the translation software to come up with appropriate suggestions, which the translator can choose to accept or decline. However, using translational software where the meaning and interpretation is subject to various views is not advisable.

Translation software can help in creating good tools such as those for checking spelling, grammar or those that suggest suitable words based on earlier translation but they cannot totally substitute manual translation. An experienced translation agency is aware of this fact. Another point to be noted is that every good translated piece has to go through a review and correction process. This is better done manually.

Some translation agencies use a combination of human skills and machine based translation.
Alternately, some agencies use a first round of automated machine based translation followed by a cleaning and refining by manual means. While translation software is also much cheaper than having a translator do the work manually, the disadvantages of machine translations are mainly related to the quality of translation. Ambiguous words having more than one meaning, improper sentence construction and wrong grammar can result in poor translation when translation software is used. Unlike human interpretation of the context and likely meaning, the software cannot use the information that is not built into the software logic. Thus tacit knowledge related to culture and usage of a language is not captured. In case there is a mistake in the original document, the translator can raise queries and get them resolved. This is not possible in the case of a software based approach.  Machine translations can only be recommended to get the overall picture or meaning of a short note or few words.

The good news is that with advances in technology and increasing need to cross linguistic barriers, more research is being conducted in the field of translation services. As newer software based tools are being created, they can be used to add value to the efforts of human translators. Ultimately, translation software and translators, both go hand in hand. Neither of the pair is indispensable, nor redundant.