anyone know of any programming models that explicitly include a first-class concept of a program’s user?

Dog does: https://www.media.mit.edu/projects/the-dog-programming-language/overview/

Interesting. Not sure if this counts, but in my visual language Hest you can (eg) select a piece of data and move it around with the keyboard like it's a video game character. So, there isn't really a single concept that represents the user, but there is the encouragement for the programmer (or the end-user programmer) to embody parts of the code & runtime data. So, in a sense, a first-class concept of the program's programmer.

@ibdknox going to do some googling tomorrow but it seems like there’s not a lot of info available on mobile at least wrt the implementation—do you know any other sources offhand? otherwise i’ll just poke around

Algodoo (the program used to make this sim) looks like it kicks so much ass. I've always been curious to try it, but haven't yet. Maybe someday.

i’m looking for something that might be strictly lamer: a language that characterizes user input in terms of its own control structure (i’m completely making up my own terms here and very open to input). so if hest lets users embody different pieces of data, the language/environment/whatever would allow the programmer to express the rules of the game for the user—which pieces of data do they get to control (and how many at a time), how can they embody them, what can they do as each piece of data (fresh thought, loose wording)

Nah, that's cool. In the case of Hest, this "play with your code/data like it's a game" is a feature of the editor, not the underlying model. I am hoping to implement some parts of the Hest editor as live data, so you can have some of that nice Smalltalk/emacs-esq self modification. But I don't yet know to what extent I'll end up doing that. It'll depend entirely on perf. Ideally, yeah, you'd be able to (eg) define some new kind of object, and then implement some special behaviour for how this "embody your object" system interacts with it. But who knows, I'm still a ways off from exploring that space.

Certainly the part of the idea in Smalltalk and Logo was for the user to imagine themselves as the object or turtle. Doesn't exactly match rise to the level of the user being an agent in the program. More along the lines of "if you want to paint fire, be fire." https://overcast.fm/+BnTo0uB-k

this podcast is persuading me to pull my oculus out of my closet to try quill again

HANDS has something similar https://john.pane.net/HANDS/HANDS.MPG. More info in: https://john.pane.net/research.html

Perhaps not quite what you are looking for, but the game “Baba is You” could be seen as an example of this. 🙂 https://hempuli.com/baba/

I played Baba Is You for the first time earlier today. It's absolutely superb.

And it keeps going and going.

both of these are very good suggestions

My Object Net is like a cyberspace where the user is a key object, that observes objects it links to, and can interact with objects it sees.

loose thoughts related to eg https://futureofcoding.slack.com/archives/C5U3SEW6A/p1567285748046200?thread_ts=1567229532.043700&channel=C5U3SEW6A&message_ts=1567285748.046200

sounds like Dreams (PS4) and Little Big Planet before that to some degree

That sounds like what I'm doing in the Object Net: there is a user response object type which is at a higher level than events. Users spawn response objects when they interact with a viewed (rendered) UI or form object.