The difficulties encountered by blind and visually impaired translators when using translation tools.
by Tara Owton
My Name is Tara Owton and I work in the translation industry as well as doing other jobs. I have had a lot of experience of using translation software such as Trados, since this was a key component of my translation studies at university. I have been blind since birth and I use a screen reader to access the computer.
The translation software that professional translators use is machine-aided human translation (MAHT.) Trados is one such software. This software enables the user to create a translation memory and terminological databases for translation purposes. A translator has a text in one language, known as the source text, and their translation is known as the target text. So using Trados as an example, the translator opens their source text, and then the text is split into segments, or small chunks of text for the user to work with. The beginning and the end of each segment are often the beginning and the end of each sentence, but this can be changed which will be discussed later. As the user starts translating each source segment into their often native language, their translation is stored in a previously created database called a translation memory. This memory enables the user to use exactly the same translated segments again. This is useful if a translator has to translate documents of a similar nature like automobile or computer manuals, or any sort of text where the language is repetitive. This increases a translator's productivity, since by using the same memory to translate similar or identical sentences, the sentences are either partially or fully translated for the user, so they do not need to translate the same whole or part of a sentence twice, meaning that they can work faster and more work can be completed in a shorter space of time.
In Trados 2007 and Trados 7, a screen reader user was able to carry out such a task. The user could use Word to translate their text, and they could move through the segments by accessing the Trados toolbar in Word which contained various different options for use when translating. Even so, there were features of Trados that were still inaccessible or very difficult to use. The concordance window, a separate window which displays possible alternatives to words when translating was visible by pressing a keystroke and the results were available in a list. However clicking on a result to select it by pressing the enter key did not work, and mouse simulation commands were not particularly effective if at all. As concordance was only an optional feature, I did not use it because it was not quick or easy to use. If I did not like a particular translation that was suggested from the memory, I navigated through the segments and edited it myself.
Other features of Trados include MultiTerm, which is a database for storing individual terms for language pairs. This and the translation capability have become less accessible if impossible to use with the release of Trados 2009. The ability to translate in Word was no longer available, so the user had to work with the Trados interface. By using the mouse simulation commands for JAWS, it was possible to see the source text, but there was nowhere visible where the user could translate the text. I tried everything, I found a list of numbers and tried clicking on them, hoping that this would enable me to type the corresponding translation, since I assumed each number represented the segments starting from the beginning of the text and continuing through to the end. However nothing worked and I could not find any way of carrying out even a basic translation task successfully with 2009. All I was able to do was set-up a translation memory.
The MultiTerm feature which was difficult but possible to use for the most part in 2007 because it was an HTML interface, became virtually unusable in version 2009. JAWS would almost constantly lose focus when trying to edit or create a term base entry, to the point it was not possible to work successfully at all. The ability to translate a webpage, thanks to the TagEditor feature in 2007, was possible and moving between the segments worked the same way as in Word. However since everything was subsequently integrated into an all-in-one package for version 2009, the ability for website translation was also lost.
The WinAlign feature, a feature allowing the user to specify when a segment should end and another begin when inserting an already translated text that is a source and target text has never been accessible in any version. This is because each segment is a graphical line which the user cannot click on to specify where they want each segment to begin.
There are other tools, Déjà Vu from Atril, Transit NXT from STAR Group, Metatexis from Metatexis software, Fluency from Western Standard, Wordfast Pro and Wordfast classic from Wordfast LLC, Catallyst from Alchemy software and Passolo from SDL (used for translating computer program interfaces and manuals), which work with varying degrees of success with JAWS. Some are extremely difficult like Déjà Vu because the navigation panes are hard to navigate with JAWS, so it takes a long time to access the particular feature you want quickly, rendering it impractical to use. Wordfast Classic on the other hand uses Word and is just like using Trados 2007 to translate. Transit NXT uses a lot of keystrokes for different tasks, making it easy for the user to translate a text. Western Standard, the company that manufactures Fluency have been interested to hear visually impaired translator's feedback regarding their software. Their software was quite useable and the user was able to do a translation.
The only tool I have explored extensively is Trados. As stated before, Trados is industry standard, so most translation agencies require it. In short, Trados is where the bulk of translation work is, and the inability to access it makes it very difficult for a translator to get regular work through an agency.
To overcome these difficulties, I have used sites such as freelancer.com to find translation, proofreading, and transcription work. These types of websites have different job categories to choose from, and I often find that the translation jobs advertised do not require the use of any translation software. Another alternative to this is server-based work. I have found an interesting accessible possibility: CloudCrowd is a company where you work through a Facebook application which is accessible and easy to use. You do the work in a Word document, and then upload it. It is these types of alternative which make some sort of career without the use of translation tools easier.
Because of the high demand for the use of Trados in professional environments, blind and visually impaired translators are currently excluded from many jobs and possibly university courses in translation. However SDL are interested in making Trados more accessible, due to requests from various sources. The latest version, 2011, is apparently more accessible with JAWS; however I have not had the opportunity to try it for myself. I am in contact with a JAWS scripter who is willing to write scripts for Trados 2011. So, provided that SDL implement accessibility features, and JAWS scripts for Trados are written, work and study prospects should improve for visually impaired translators.
It could be argued that if everyone learnt JAWS, Window Eyes, System Access or NVDA scripting depending on the screen reader they use, people could write their own scripts, thus making everything they want accessible. But learning scripting takes time and patience, and is not easy if you have no programming background, or if you find programming concepts difficult to grasp. Scripting is also costly, particularly for JAWS which is the most used screen reader in the workplace. The ideal solution would be for translation software, notably Trados, to be accessible with screen readers when installed, no scripting required. That way people would not have to rely on the fact that there are scripts for only 1 particular screen reader available, and said scripts may not be available for the screen reader they are accustomed to.
If scripting and accessibility features are not implemented, then blind and visually impaired translators will continue to be excluded from the majority of available work. Alternatives such as accessible cloud-based working environments are possibilities for the future.
These issues are often discussed on a list for blind translators and interpreters.
http://lists.screenreview.org/listinfo.cgi/theroundtable-screenreview.org