Follow

ummmmmm i just saw bret victor talk about dynamicland and it upended all of my conceptions about programming, how had i not even known about this until now dynamicland.org/

@aparrish beautiful! I only started in 2018 apparently, so don’t be too hard on yourself!

@aparrish
It's come up a couple times over the past ~year in UI discussions I've been in on here (since Bret Victor is a fixture of the whole "it's theoretically possible for GUIs to not be shit" community).

Details of implementation & use seem to be thin on the ground. Did the lecture you attended have those?

@enkiv2 no implementation details, sadly. the "use" seems mostly to be "let's play around with the affordances of the system we built" rather than "here's a particular problem we're trying to solve" (which is very much okay with me, since I think the affordances of the system are very interesting)

@aparrish
By 'use' I meant examples of interaction (which might highlight stuff about implementation), rather than application per-se. (I also think the system looks super interesting to play with, but I've been unable to figure out how you interact with it beyond "there are cameras and projectors, and some of the cameras can read some kind of environmental visual code" but it's unclear if that's mostly QR codes or what)

@enkiv2 oh! there was lots of that. each sheet is an object that has its own code written on it. (not sure if this is "actually" OCRed from the page itself or stored in a central DB). as I understand it, objects make claims about their state & what they want to happen, then other objects can act on those claims. the interesting part is that objects know where they are in space, so you can say "make whatever is left of me turn pink" or "get the value for this variable from the object below me" &c

@aparrish
That's good to know!

It looked like the pages were too far away from the cameras to be OCRed, which makes me wonder how they were even identified.

@enkiv2 if you look at the recent demos, there's are large colorful dots on the border of each page, which is how they're (currently) identified. but tbh I think a halfway decent cv model could be trained to recognize them with much less

Sign in to participate in the conversation
Mastodon

Follow friends and discover new ones. Publish anything you want: links, pictures, text, video. This server is run by the main developers of the Mastodon project. Everyone is welcome as long as you follow our code of conduct!