Sunday, November 26, 2017

"Syukuran" dua tahun di Inkubator Duolingo

//reposted from https://incubator.duolingo.com/courses/id/en/status

It's Thanksgiving weekend here in the United States and I thought it would be fitting to reflect on the progress that we've made since Meika wrote an update a year ago.

The main progress update that I can share with you is this: we now have the lesson tree completely mapped out. There will be ~1900 words that we will teach in 69 skills. We will continue to refine the word list as we add them to the course but now that we've built a map for the lesson tree, we should be able to progress steadily.

For me personally, this is an important milestone. We started this course nearly two years ago (Dec 2nd, 2015!) with the idea that we should prioritise teaching useful and frequently occurring Indonesian words. However, we didn't have a clear metric to select words from the Indonesian vocabulary to be included. I tried using the Indonesian corpus from OpenSubtitles repository and different online vocabulary lists, but I had never really felt satisfied with any of them. What I finally found satisfactory was the Indonesian word frequency list from Ivan Lanin, Jim Geovedi and Wicak Soegijoko's 2013 paper who have made their data available on github. I decided to err on the side of formality and use their Kompas and Wikipedia corpus as they require less additional data processing (as compared to Twitter and Kaskus which promised headache in trying to sort the standard Indonesian spelling from the colloquial ones). While the composite still require some sorting, this was a huge leap in completing the first roadmap for the course.

As I mentioned above, we are continuing to refine the plan as we build the course, and it is for this purpose that we have recruited Joe (@jsullysull), a UK-based graduate student to try out this course as we're building it. He has been sharing his impressions on the course and so far his reports have highlighted the need to write clear tips and notes for the skills for future learners (I personally used Duolingo more often on my phone so the importance of this has largely escaped me).

Nevertheless, our focus for now remains to add the words from our lesson plan to the incubator and create sentences for those. I have recently completed making the "Me-kan" skills that will teach causative verbs, and the next two skills on my list are "Education" and "Comparison and Superlatives". That being said, if you have been following our progress from the WIU update, you may see that the progress bar for this course will decrease/remain stagnant as we add new words to the course (boosting the progress bar requires making sentences to complete the skill, but making sentences require more thought and will of course be slower!).

Lastly, I'd like to give a shout out to contributors who was active in the past year: Meika (@Maycca), Ahdiat (@Ahdiat-P), and Nindita (@NHarahap). Thank you for your contributions to this course.

Until next time—tabik!

No comments: