As a company specialising in website and web-based translation, Web-Translations likes to keep its finger on the pulse when it comes to translation technology. This month we upgraded to Trados Studio 2017; here are some of the new features we’ve experimented with since the move, along with some of our thoughts about their relevance to our day-to-day workload.
The first thing we noticed when firing up Studio 2017 was the changes in aesthetics. Whilst we were relieved to see that the basic user interface has changed very little, we did notice some brighter hues, as well as a new colour-coding system on file completion bars, which is a nice touch.
Another feature that immediately caught our eye was the option to drag and drop files for translation; this is an amendment to the import process for single file translations, and it does speed things up somewhat. As Project Managers we usually work within established projects with multiple files, but we can imagine this feature being useful for freelancers working with a greater number of single file jobs.
One of the biggest challenges we’ve found over the years has been the inability to merge source segments in Trados. Sometimes, for reasons beyond our control, a source text will import in such a way that sentences become split over two or three segments; this causes a headache for our linguists, and leaves us with TM entries that only have limited use. It’d be interesting to see how the merge features compare across MemoQ and Studio 2017 in terms of functionality and ease of use, however, and also whether either offers a more efficient means of merging for larger file types.
Fuzzy match repair is a part of Studio’s new upLIFT technology. Again, we already work with fuzzy match repair systems in other CAT tools, and don’t yet find them to be effective enough to trust in every instance. A comparison between Studio’s fuzzy repair and other CAT tool equivalents may be beneficial here, before deciding on our system of choice.
Sub-segment (fragment) recall is another upLIFT feature which has also been available with some of Studio’s competitors now for a few years; SDL have, until now, left users to opt between Autosuggest or concordance searches, both of which leave some room for improvement. upLIFT fragment recall works by breaking up your existing TM into smaller translation units, which are then indexed in the TM for quick look up. Translation units can be as small as you like, meaning that TMs will then have accurate recall on even two or three word phrases within a segment.
We can think of several client accounts that this would come in useful for, and see this is as an interesting streamlining of two resources: a standard, traditional TM and a basic termbase (without context notes).
Studio’s AdaptiveMT feature is essentially a personal machine translation tool that learns from your segment edits. As a company, we don’t currently use MT in a commercial setting, and as such may not make much direct use of this feature; it may prove more useful for companies (and linguists) working predominantly with a large volume of technical documents; creative texts are always going to be too specific for MT!
We’ve found a few changes up front within Studio 2017 that will have a direct impact on our productivity and workflow, and we can’t wait to experiment further with some of the more complex features such as upLIFT. As a company, we’re not entirely sold on AdaptiveMT as it does not really suit our workflow; we remain very much in favour of skilled human translation. That being said, in an industry at the forefront of technological advances, it is really important to keep an eye on developments, so as to always be making an informed decision.