Episode 1: Ship Fast, Translate Faster with GitLab's María José Salmerón Ibáñez скачать в хорошем качестве

Episode 1: Ship Fast, Translate Faster with GitLab's María José Salmerón Ibáñez

If your coding workflow is built around continuous integration and continuous development, like GitLab's, you can’t afford to wait for translation. As GitLab's Senior Localization Project Manager, María José Salmerón Ibáñez, explains in the first episode of DeepL’s “The New Fluency” podcast, it’s time to design for continuous localization. In this fascinating first episode, María José talks to DeepL's Director of Localization, Morana Perić, about: How to adapt translation to the ways that technical writers communicate Where automation and integration is lacking in translation systems The imperative to ensure that every member of a developer community can contribute Why localization teams should embrace linguistic debate Here are some of the highlights: “I think we are living in a really great time. AI is giving us a lot of opportunity and a lot of room for explorations. AI brings the benefit of like doing fast translations, and we are combining this with the traditional benefits of having a TM-based localization workflow. We are trying to combine and find the best recipe with how can AI bring us the speed that the DevOps industry requires us to do with all adjunct development. We need to, uh, deploy translation fast and accurately with the benefits of having a traditional translation, memory-based workflow and TMS-based workflow. So for, for example, for GitLab's technical documentation, we really need to understand how the content is being processed and how English is being processed in the, in the source, how technical writers are creating the content. We can then iterate over the content so that we could design the best AI-based solution that would really plug into the system without being disruptive. In a GIT-based system, we don’t just require AI translation. We also need to understand what GIT is and what are the requirements and how are developers and how are technical writers working around the content, and how can the localization team really prove and show the value of, um, plugging in localization into their systems.” “In the localization industry, we are used to this waterfall: first A, then B, then we cannot start translation until the source content is finished. But the AI is giving us this power, as well, of iterating on the translation fast. So you don't really need to wait until the English is done to kick off translation. Also, when you work in, GIT-based system like GitLab, with the way the continuous integration and continuous development, so the CICD pillars that the code is being built on, you also need to embrace this and design how localization can be continuous. What is true continuous localization and how can I use the tools that I have to make this agile workflow tool? It's a very nice, challenging thing to think about, because not every tool is designed for these agile environments.” “This whole traditional localization tooling that’s built for enterprise scale - sometimes the automation and integration that they provide is not enough. You need to do more automation and more integration. You need to really do the pre-translation work and internationalization work to prepare everything before you actually kick off translation. That’s the blocker, but also the opportunity.” “The community contributions that GitLab has are really inspiring. It’s also a challenge for the localization team, because this philosophy of everybody can contribute, we have to ask: what does it mean for localization? Because they can also contribute with translation. So how, what do we allow people to contribute? How do we integrate these contributions into our workflow and into our translation memories and, frameworks, and how do we manage these community contributions? It’s a really beautiful thing to see and to manage and to allow. And I think it's also something that AI enables because, and this goes back as well to the whole agile framework or agile methodology, it’s not only linguists or AI that are the ones producing the translation. What happens when developers or marketing people, or other people that just love languages, want to contribute to translation? There are a lot of linguistic debates opening and a lot of like exchanges and conversations and also technical challenges of how to integrate these community contributions into the, the localization tech stack and translation memories.” #TheNewFluency #LocalizationExperts #LocalizationWorkflows #LanguageSolved