It feels like taking a picture and converting it to the lowest image format. Even if you know what was there previously, it's impossible to convert it back to a meaningful piece of information.
Trying to interpret sentences out of context in Toki Pona is indeed tricky!
However, "sina wile ala wile moku e telo pimeja?" (Do you want to eat black water?) can be quite accurately reconstructed as "do you want a coffee?" if you know it was said outside of a cafe.
With a given context that you can refer to (in a cafe, walking in the park, cooking food), the expressive power/range of reference of 120 words is vastly greater than, say, in a letter to a stranger. This is probably one reason why the meme community in Toki Pona is reasonably lively ( https://www.reddit.com/r/mi_lon/ ).
I don't know how you extend that metaphor to image-decompression though! (I guess if you know you're decompressing a picture of a face you can train your decompresser on other picture of faces first :) ).
having played with it i will say that it was an interesting and valuable experience to try and distill something that's emotional to you down into the basic basic concepts required to try and express it, which is exactly the point of TP. It's not for trying to have real conversations in.
However, "sina wile ala wile moku e telo pimeja?" (Do you want to eat black water?) can be quite accurately reconstructed as "do you want a coffee?" if you know it was said outside of a cafe.
With a given context that you can refer to (in a cafe, walking in the park, cooking food), the expressive power/range of reference of 120 words is vastly greater than, say, in a letter to a stranger. This is probably one reason why the meme community in Toki Pona is reasonably lively ( https://www.reddit.com/r/mi_lon/ ).
I don't know how you extend that metaphor to image-decompression though! (I guess if you know you're decompressing a picture of a face you can train your decompresser on other picture of faces first :) ).