Patrik Bulak Shares How CAT Tools Work

Patrik Bulak Shares How CAT Tools Work

CAT tools specialist Patrik Bulak has been part of the Skrivanek staff for nine years this September. We asked him some questions about this work, his career arc, and his personal life, and learned a lot of interesting details about him and the way CAT tools function.

When and where did you begin your career, Patrik?

During my studies of the Czech language, Czech literature, and general linguistics at the Faculty of Arts of Charles University, I specialized in cognitive-cultural linguistics. In this context, as a research assistant at the department for the study and analysis of the Czech Language at the Faculty of Arts, Charles University, from 2012 to 2014, I actively participated in the organization of conferences, co-edited professional publications, and did some teaching.

When did you join Skrivanek, and in what role?

I started my career at Skrivanek as an in-house checker in September 2013 while maintaining my faculty duties. Thanks to this experience I was performing output accuracy control of delivered translations in all possible language combinations. An in-house checker is the last pair of eyes to examine a document for errors before that translation is delivered to the client, an aspect of the workflow that can also done by the project manager. This is basically the final proofreading before we pass the document off to the client, and it’s required for all translated texts whether or not the client has specifically ordered proofreading. The kind of errors at this stage that we are looking closely for are, for example, changes in formatting that don’t correspond to the original formatting, accurate use of numbers, and spell-check inaccuracies.

Has your role/title there changed?

Within a few months I became a senior in-house checker, training and mentoring a rapidly growing team of new in-house checkers at the local back office, mostly students of linguistics just like me. As in-house checkers we worked alongside the project managers at the local back office, which was responsible for purchasing and management of professional services.

After finishing my studies in September of 2015, I was offered a full-time position as a project manager at the local back office. I thoroughly enjoyed this position. I enjoyed the communication with our suppliers, but the technical side of project management was even more interesting to me. In that position I was importing documents into CAT and preparing them for processing, sometimes also handling the linguistic aspect of processing the documents. I found this to be a particularly interesting job and I had a lot of responsibility.

It was then that I started to think more deeply about the translation memories of our clients, and also the advanced possibilities of terminology databases. For instance, I could see ways of building glossaries based on the most frequently used words right as the translator worked. When I received the offer to work in the CAT department in April 2018, I didn’t hesitate for a second.

When did you start learning about CAT tools, and what was that like for you?

I’ve always been closer to the world of software than hardware, so CAT tools were already a fascination for me as an in-house checker. CAT tools, thanks to the built-in QA (quality assurance) checker, make it possible to effectively check about twice as many words (sometimes more) by sorting potential errors into key categories rather than having the human translator simply eyeball the translation against the source document. Again, examples of key categories in this sorting process are formatting, accurate use of numbers, spell-check and even terminology adherence, and untranslated parts. It’s a marvelous feature that CAT tools then allow the document to be safely exported in its original format, and thanks to the most advanced CAT features it is no longer necessary to humanly check everything after the export, as most of potential errors were repaired directly during the CAT tool’s QA. The trickiest part is when we have to deal with more complicated formatting or when we are faced with express order processing.

How have CAT tools changed/improved since you started?

The evolution of CAT tools follows the trajectory of developers trying to increase processing efficiency by adding new features and expanding the portfolio of importable formats. But there is also a trend toward offering individual feature options to more demanding clients who choose to manage projects themselves with only select Skrivanek services incorporated into the translation process.

Of course, another improvement is that neural machine translation (NMT) possibilities have expanded recently, and we responded to this with Skrivanek NMT. Thanks to this shift, the efficiency of order processing has again increased significantly, with translators moving in recent years from full-document translation to post-editing of raw NMT translations.

For a layperson who has never seen or used CAT tools, can you explain how they work?

CAT means computer-assisted translation. As the name suggests, the human user works all by himself and the computer is only an assistant. The CAT’s functions are based on translation memory created by the user. When working in CAT, the source text is divided into separated segments (usually sentences), which the translator translates and the CAT tool stores in the translation memory. It’s stored in this form: source text + target text (the translation). When/if the stored segment from the source sentence appears again in the text, the CAT tool’s translation memory offers the translator the translation of the last-used version of the phrase. If a similar segment appears in the text, the translator is offered a partial translation to use (called a fuzzy match because the match is not 100%).

The CAT tool’s translation memory is used to draw on translation units stored in the past, while simultaneously storing translation units for future translations as the translator works through a client’s documents. CAT tools thus lead to time and financial savings, but even more importantly to an increase in translation quality (through consistency, for instance). Last but not least, CAT tools allow the import of formats that it would otherwise not be possible to translate directly (e.g. web formats etc.).

Patrik and the team

What do you enjoy about working with CAT tools?

As cliché as it may sound, I enjoy how CAT tools are fundamentally identical due to a common principle, and yet so different in some ways. For instance, some are more user-friendly, others less intuitive, and some so sophisticated that it can take an inexperienced user a while to employ them to their full potential. The compatibility of the various CAT tools has improved over the years, which makes the work even more enjoyable. It is also interesting to see trends where one CAT is “catching up with the others”. For example, at the beginning SDL Trados Studio only enabled work offline with sharing of so-called project and return packages; now we can work online and multiple users can see progress on the project in real time. I really enjoy the situation of many possibilities that CAT tools create.

Would you give an example of a project that would be extremely tedious or expensive without the assistance of CAT tools?

A good example is the type of project where our task is to translate technical texts in which 90% of the total word volume is repetition. Individually translating such a high volume of repetitions would not only be inefficient in terms of our processing time, but also unnecessarily expensive for the client and more likely to produce inconsistencies (which will then have to be solved during QA at the latest), especially for larger word volumes.

What sort of problems can arise with these tools?

Sometimes a client has already translated documents from the past and wants to take advantage of the translation memory in CAT. If both the source and target documents are the same we can create a translation memory for a new project quite easily and offer the client a discount. However, if the client’s translation differs from the source document in terms of content and only some passages match, creating a memory is very problematic. Often it is better to offer the client a frequency analysis of his source texts, from which we export a list of the most frequent words and then we can have these translated by the selected supplier and, after both parties have agreed, the terminology database thus created can be used in the project.

How would you summarize the role of translation in the world and how are CAT tools affecting that role?

The role of translation, in my opinion, is the transfer of information. With the increasing pace in today’s world, demands are naturally placed on speed, but at the same time we strive to maintain the quality of the product, ideally at the lowest possible cost. So while a few years ago the CAT tool was an interesting option, today it is more and more commonplace because it does all of these things. A manifestation of this “speeding up” trend in our dynamic industry is the use of other translation technologies in the processing of projects; MT is an example, the quality of which is increasing more and more not only every year, but literally every month.

Please tell us a little bit about your life outside Skrivanek, Patrik.

I live in Prague with my wonderful family, my wife and a son who will be 3 fairly soon. We have an almost 8-year-old parakeet living in the apartment with us.

Thanks to my parents’ roots, but also to my studies at the Faculty of Arts of Charles University, I especially like the Slovak and Polish languages. However, my main foreign language is English.

Music has been an important part of my life since childhood. While at work every day at Skrivanek I am immersed in CAT software, but outside of it I am a DAW (digital audio workstation) software enthusiast. Thanks to this home-studio workflow, I can create my music tracks completely by myself from drums to vocals (I especially enjoy the mixing and mastering phase). But apart from that I am the drummer for a real-life alternative rock band, and I think this can be considered my main sport. I also like riding my bike, but since all of these activities take place in a seated position, I try to at least walk a lot, and I practice yoga, Tabata training, and breathing exercises.

Skrivanek has had a big presence in your life for eight years – can you tell us one characteristic of the company that has contributed to your love of your job?

I think it’s all about the atmosphere Skrivanek creates. When I look back over the years, for me it’s about the opportunity for continuous individual growth and equally about connecting with others on a surprisingly friendly level that goes beyond our work interactions.