First, I had to log in to the Google Cloud Console using my Google account. Then, as I learned from the webpage https://fotc.com/pl/blog/google-translate-api-czyli-jak-pobrac-i-zainstalowac-klucz-translatora-google/, I needed to create a project and enable billing, as there is a limit on the number of free characters I can use in my translations. Once this limit is exceeded, usage fees are charged for using the API key.

After creating a new project in OmegaT, setting the source and target languages, and adding the source text to the project, I began the installation of Google Translate. To do this, I navigated to the Tools tab and selected Preferences. In the displayed window, I enabled Google Translate v2, entered the previously generated API key, and then secured it with a password.

To my surprise, after completing these steps, machine translation was still unavailable. None of the solutions I found online helped me resolve the issue. However, after some time, when I took another look at the program, I noticed an unchecked box labeled Automatically fetch translations, which I had previously overlooked. Once I enabled this option, machine translation started working. This part of the process definitely made me realize that patience and attention to detail are essential when learning to use CAT tools.

Google Translate provides suggestions for each source text segment. To accept them, I've used the keyboard shortcut Ctrl + M which I found in the OmegaT user manual: https://omegat.sourceforge.io/manual-latest/en/appendix.shortcut.custom.html

The translations suggested by Google Translated turned out to be not entirely accurate, so the final version required significant revisions, which was to be expected in the case of a text that demands a certain level of creativity.

From multiple online sources, I learned that it is important to secure the API key to ensure safe usage. Thanks to the guide below, I was able to restrict the key to Cloud Translation API, as shown in the screenshot above.

https://www.youtube.com/watch?v=2_HZObVbe-g&t=81s

Summarizing my self-learning process, I can conclude that this exercise not only helped me discover a useful feature that facilitates translation in the CAT tool OmegaT, but also allowed me to learn how Google Cloud Console works and how important it is to secure my keys and projects. I also learned that when using the Google Translate engine while translating in OmegaT, it's important to pay attention to potential errors in the content, such as: inconsistencies in gender endings in Polish, incorrect word choices, illogical phrases that may be a result of calques or machine translation errors.

As for what I would like to learn in the future, I am definitely interested in using multiple machine translation engines in OmegaT, as this could help in selecting better translation solutions, as those suggested by Google Translate were not always perfect.

I would also like to try activating Google Machine Translation in another CAT tool, such as memoQ, which I find much easier to use due to its more intuitive interface, although this is unfortunately a paid tool. Nevertheless, OmegaT performs quite well for a free CAT tool and can be a good starting point for beginners without a sufficient budget to purchase licenses for other tools such as memoQ or Trados.

In my self-learning project, I aimed to explore and integrate machine translation (MT) into my workflow using the CAT tool OmegaT. Specifically, I focused on configuring and utilizing the Google Translate engine within OmegaT to enhance translation efficiency. The choice of this topic stemmed from advice given by my course professor, which I found very interesting. I decided it was definitely worth trying out, as machine translation can undoubtedly speed up the translation process.

As the source text I used a fragment of a text about "Flaking" (included in the attachment) which was translated during translation workshops, to see to observe how Google Translate handles problematic texts. The text was challenging, as it contained numerous expressions that are difficult to translate into Polish in a smooth and correct way. This includes, for example, the title concept itself, as "flaking" lacks a direct and natural equivalent in Polish.