Hacker News new | ask | show | jobs
by voisin 2165 days ago
Rather than collate yoga videos with differing levels of accuracy relative to search (the search I just did was pretty inaccurate), why not have some gifs of poses and generate the practice by linking these poses together to mix something based on the user’s chosen level of difficulty, style, and goals?

For my example, I chose “Advanced” and two of the three videos were labelled “Beginner”. The third I did a quick view of and it also looked like a beginner practice. This is an issue generally with yoga videos online - they are predominantly for beginners with somewhat limited Intermediate videos and very rarely Advanced videos. But you could solve this with gifs of different poses linked together on the fly, and generate the exact desired length of practice!

3 comments

That's the approach that the Down Dog yoga app seems to take. They had an instructor go through all the poses and transitions, and they can then run all sorts of different classes by combining them in different ways. They do it pretty seamlessly, but i have to imagine a lot of planning went into it.
Interesting I'll check this out. I had a similar idea that you would basically catalog everything and then have instructors put in classes. And as you grew more and more class you could start to find repetitions that instructors use and eventually be able to randomly generate classes with well known sequences.
Thanks I will check it out
This is tricky because practices aren't just random sets of poses. I don't know much about yoga sequencing - maybe you could generate a reasonable set of rules as to how to combine poses but it doesn't feel trivial to me. Also 'beginner' is hard to define - familiarity with poses is independent of the progression within them which one can achieve!
Yes, but also include spoken (audible) instructions.

It's not always easy to look at a screen while in yoga positions.