Crowdsourcing and Translation

Crowdsourcing has been a topic of hype for a number of years in the translation industry, but the reality of it will shape the future of the translation industry. So, does crowdsourcing translation really work? and how does one guarantee quality and consistency? The main concern for companies is entrusting their strong corporate brands and style to unknown volunteers, but the bigger issue is that people are less likely to offer translations for corporate materials in which they hold no stake, and have no emotional investment. For instance, it's unlikely that companies could crowdsource the translation of packaging and labeling information.

The benefits of crowdsourcing lie in in the world of social networking. Facebook's experience with crowdsourcing has shown that they can break down language barriers with the help of their users. Thus making it a stepping stone showing that for-Profit companies can leverage the enthusiasm and language skills of their user base to achieve their business goals. Start-up small businesses are increasingly venturing into social networking by creating communities that are passionate about the product, and willing to translate ads and promotional material. Thus, providing the companies with a wider audience, where they would have a real interest in feeding into the translation process.

The viability of crowdsourcing is giving rise to translation companies like myGengo that are structured to offer clients the best of both worlds: low rates by taking advantage of crowdsourced translations, and some level of quality by reviewing the translations. There are no guarantees on intellectual property protection and consistency, but it's a cheap/fast way of getting translations done. So, is crowdsourcing translations the future of our industry? Probably not. Will crowdsourcing play a role in some sectors of the translation industry? Probably.

Where are we headed? Integrated and Automated Translations

With technology rapidly changing and our need for instant gratification, the translation world as slowly evolved over the years to meet the demands of our fast paced life. The advent of CMS (Content Management Systems) has facilitated this process. It has allowed us to move content from one source to another with email exchanges and zipping files. However, these systems are not cheap, and smaller translations companies are finding hard to understand the integration aspect when a client says: "Can you work with our CMS?"

The complexities vary from one CMS to another, but the clients expectations are often the hardest for translation companies to meet. The integration process is usually the easiest part, if the client is willing to allow the translation company to access their CMS and "grab" approved content for translation. There are also integration technologies like "Global Link" by Translations.com that allow a "seamless" integration between Interwoven a popular CMS that is capable of dealing with translation and localization technologies and SDL Trados.

The bottom line is the amount of work and cost involved for both sides. Current FTP technology and Web based servers can create workflows that allow clients to export content form their CMS and get it to an FTP, and a notification is sent to the translation company, who then downloads the files. The translation company reverses the process for the translation (either through their CMS, or uploading files to their FTP). In this scenario most of the work is handled by the translation company. Overall, integration technologies are common palace and it's more about choosing he right fit then the integration process itself. The challenging part lies in having a editing module where clients can access the translation memory and get translated content as the create or update documentation. A simple integration/file transfer will not accommodate this process. Always test out new technologies first, most companies offer Free Trials, so it's a no brainer.

Multilingual or Not Multilingual, That is The Question? Right

Two weeks ago when it was ridiculously cold in Milwaukee, I had to jump start my car so I could get to work. While in the process of trying to jump start my car, the car's alarm decided to go off , and I frantically tried to disarm it. I could not find a way to disarm it, and I had to consult the owners manual. 8 minutes of deafening noise and evil-eyed neighbors later, I finally managed to disarm the alarm with a few coordinated clicks of the key less entry remote. In those 8 minutes of flipping through the owner's manual of my car, I was shocked to find that the bilingual manual was laid out in weird line by line sections without proper page breaks. So, as I continued to read a sentence at the end of one page and instead of continuing on the top of the next page, it continued to the middle of the next page, while the top section listed the instructions for operating the keyless entry for the trunk. This bizarre bilingual layout got me thinking about the best way to have multilingual layouts for manuals.

I think part of the problem was not planning ahead for a multilingual manual. The cover and index page were in English and since it was for the North American market and specifically the U.S., they decided to place Spanish translations below each English instruction and ignored the re-flow of text and page breaks causing text to jump from one section to another without re-adjusting the images that correspond to the instructions. I think the best way to create a multilingual layout is to have sections for each language. I have noticed that most manuals have this type of layout. The car company in question did revise the layout of the owners manual for the next model year to a trilingual sectioned setup. It looks much better, and it's easier to locate information. By the way, my neighbors still give me evil looks. Once again to those affected, I apologize for the noise.

Unwanted Files

The horror of digging through files that are not required for translation and localization has made the idea of driving of the bridge seem like a dream. I understand that sometimes people don't really know what files they are sending, they request it from a department in their company, that dude zips everything and forwards it to you; you send it of to us, and it's a mess.

This is particularly true for websites. Websites are usually composed of various files that are supplied by different groups and compiled by the webmaster. So, when it is sent for translation, the webmaster usually sends the entire package without indicating what needs translation. Usually, html, flash, xml, and graphic files are translated. PDFs that relate to another document are not translated unless specified. CSS and RESX files are excluded because they do not hold any literary content. Well, please try to specify what needs translation when you send a bunch of files. Accurate navigation will help the translation process, and keep my car from crashing through the guard rails.

Fonts Fun

Desktop Publishers will sympathize with me on this topic. Fonts are a big problem when translated materials are desktop published. It is important to find out what platform the original document was created on: MAC or Windows, and then if there are compatible fonts for that language. Most Western European languages have fonts that are built into most of the software programs, and are compatible with both platforms. True Type Fonts (TTF) are a universal font type that you can install on both MAC and Windows. There are certain groups of fonts that are specific for Adobe's software, checkout the list at: Adobe Fonts

Asian languages have various fonts that are specific to each language based on the character type. For example: PMing Liu (包含鋰離子電池) for Traditional Chinese, and Sim Sun (包含锂离子电池) for Simplified Chinese. They may look similar to the common eye, but an expert desktop publisher will know that there is a difference, and you don't want your readers reading the wrong thing. Therefore, you can have your content translated correctly, but if you use the wrong fonts then it's a complete waste of time and money. On a side note, as I was writing this post, a client sent me an email asking what fonts I used for Arabic, Chinese, Polish, and Russian. What a coincidence since I was blogging about the importance of knowing and having the correct fonts for translated content.

Straight to the Point

Technical Communicators know the importance of getting to the point. The instructions they write are straight forward and easy to understand. When it comes time to translate, a good translator will not expand or embellish the text. They will translate it using the most concise terminology available. European languages tend to expand when they are translated from English, so a good translator will know how to make it as straight forward and understandable as possible. This maintains the integrity of the English content, and the original intent of the Technical Communicator.

Paper Out, Websites In

During lunch today, my colleagues and I had a discussion about the end of printed materials. Will magazine printing and book publications finally come to an end? Will we stop reading printed copies of newspapers? Well, almost every publication has an online subscription, and most major news organizations have reports and videos on the web. It's a natural progression from print media to online reading or viewing. I used to get subscriptions for auto magazines (yes, I'm a big car geek), but now I just read blogs and edmunds.com for the latest news. Occasionally, I would print an article or two if I wanted to read it at the gym, or if I have to wait in line at the DMV (hell on earth). I think it's safe to say that printed publications may come to an end in the near future, but what does that mean for translated content?

Many websites are translated, and most marketing content that appears on the web is translated. As the print media slowly dies off, the shift to web based formats will see steady birth rate. This will change the game for translation companies as they have to adapt to on-demand or live translations as content gets updated frequently, and there is no boundary to the 24 hour news cycle. Some companies have employed machine translations that are linked to an existing term base to generate professionally translated terms and ensure a certain level of consistency. It may work for smaller volumes of non-scientific or non-mechanical content, but the lack of a human translator in those fields will cause major terminology errors. The resulting effect will devastate a clients reputation. The move to CMS and database driven websites eases the translation and localization process, but a 24hour on-call human translator is still a few years away. The hybrid alternative is having a machine translation do most of the work, and a translator or editor review and edit the content. This is a little time consuming, but it will have a better result than purely relying on machine translations.

Only time will tell how far we move away from print media, and how much of the internet will consume us all. In the mean time, translation companies are starting to adopt online management tools to handle content. Server based technology has come along way with XML content and DITA constantly evolving, we might have a shot at a purely online world.