I hear https://dynamicland.org/ will update sometime today..

Early bird gets the worm!

oh the preview image is already from the new site

its time

🎄

It makes me mourn all the things I didn't document at all

A few snippets I've enjoyed from the new site:

I loved that it was just all over email. Very simple.

Love the polyglot vibe. (https://dynamicland.org/2022/Progress_report/)

A few reactions. The crux of the argument seems to be that it's hard to foster literacy in dynamic media (and therefore systems thinking at population scale?) with small individual-scale screens that tend to keep people learning separately. It'll be easier to accomplish if we can learn together, experiencing computational media together in the space around us. It's not the thing, it's just the thing that will get us to the thing. I sure hope we get to the thing. --- One cost of physical things: They take up space. This is something I'm acutely conscious of as I sit in my tiny home in San Francisco watching videos of huge rooms containing dozens of posters -- with no affordance for scrolling. To the extent that the goal is to help cope with say climate change (I can't link to 6:10 in https://dynamicland.org/2024/Intro), this seems like a bet that low-density places will be more competitive in the next 100 years. Even as the habitable area on Earth shrinks. To be fair, I do have tons of unused wall space high up. There's also likely lots of room for such collective spaces to interplay with more private screens. --- It's unfortunate that I can't search through almost all of the site. Again, this may be ok. Their primary goal is for the site to be physical. The web version is second class, and that's a reasonable opinionated position to take.

> One cost of physical things: They take up space. I've been thinking about this a bit today. I've seen a few people bring this up. I think a lot of the way that DL currently "uses space" is an incidental complexity — just like the cameras and projectors are an incidental complexity. It seems like they want dynamic objects, not necessarily big posters and binders and tables full of stuff. Like, I think it'd be perfectly reasonable to assume they'd like you to (eg) be able to buy any old pocket notebook, and for every page in that notebook to be augmented by itself, not needing any extra hardware. They talk a bit about a change that took place when they set up the bio lab — they got a higher-res projector and upgraded their recognizers, and this let them make their tracking dots smaller, or do away with the dots entirely, so they could use new kinda of objects as dynamic objects. So I think it's safe to assume that they want that trend to continue. So in the long run, there's no difference between the space you already live in, and the objects in it… and the dynamic version of that space and objects.

I'm a physical minimalist but a digital packrat. The reason for the dichotomy: it seems much easier to not make messes in digital spaces. Physical spaces will be messy, particularly collective physical spaces. Doubly particularly, collective physical spaces with kids. So to me all this looks nice in the way a curated shot looks, but I wonder how much it'll hinder learning when you can't find all the space to lay out the next video. (This is a different point from things looking handmade, which I don't have a problem with at all.)

it seems much easier to not make messes in digital spaces

I dunno - my computer desktop is always a mess 😅 Also, without intentional organising, second brains can be a mess (and often abandonded). I find the infiniteness of digital being a burden sometimes, whereas with physical things/spaces, it can be more human scale and manageable. Marie Kondo taught me that being organised is more a mindset 😛

@Alex McLean it's part of the beauty of live coding. i think you told me that recently :)

I think it's ok to regret beauty 😉

Wow the projects page only goes to 2016 - I guess it might be a while before the newest stuff is added

Pages after that exist, they're just stubs. Eg, this page about weaving, that I'm super duper interested in (and should probably ask @Alex McLean about) — https://dynamicland.org/archive/2024/Weaving

heh that's us playing with the fibonacci sequence, which Geraldine uses in her cycloid weaving https://forcesintranslation.org/cycloid-weaving-spirals-and-a-few-knots-a-link-to-an-online-introduction/

Forces in Translation seems like a total jam. I'm going to enjoy sharing this with my partner (who helps run a weaving guild)

This particular activity really speaks to me: https://dynamicland.org/archive/2024/Archive_editing

Why dynamicland/realtalk is not open-source:

This one is really pretty: https://dynamicland.org/archive/2018/RGB_Color_Book

The comment about open-sourcing points to a problem that I see happening in computational science: a growing divide between two factions with opposing views, one saying "all scientists need to learn software development and its tools", and the other one saying "let us do our science and give us tools that are actually appropriate for that". Needless to say, it's the first tribe that develops all the software, while the second tribe either stays away from computing, or runs code without understanding what it does.

@Konrad Hinsen 💯 And this is why LLMs are so promising. They can help make machines more human-like instead of forcing humans to speak the machine's languages.

@Nilesh Trivedi Time will tell... Science requires high-precision communication, between humans and with machines. I wouldn't trust today's LLMs to be reliable enough. But it's still early days.

The best part of https://dynamicland.org/2024/Is_Realtalk_open_source is pointing out programs may not have source code, something I still don't understand. The worst part is insisting collaborators be physically in the same place. I hope it's building up to a point where I can interact intensely with dynamic objects in your office and vice versa. I particularly like the offhand suggestion that "community of practice" = "teachers teaching teachers".

Which always leads me to wonder why the Desktop is the only folder that remembers locations.

This got lost in the transition from OS 9 to OS X. In the classic era, Finder would remember the window position and display settings for every folder, in particular the arrangement of icons (for whenever you switched to icon view). If you put something somewhere, that's where it'd be henceforth. OS X kinda does this sometimes. It's just a tendency, not a promise. Never forget what they took from us!

Minimize your dependencies, Luke. Not just dependencies for your software. Life dependencies. Here's the topology of my major[1] backup automations. Everything funnels down to one of three places for the most part. Mostly I just search in one place in my laptop. But I hear what you're saying. Everybody isn't wired to do this. Even digital spaces will struggle when they host collectives. I give up something by being intensely alone with my data, by prioritizing the ability to export over the ability to work with others. [1] Doesn't show minor details like the periodic manually triggered FoC backups that funnel into X13.

Here's my understanding of what's going on with the Realtalk source (based on scattered readings over the past day). At one point (2019-ish), Realtalk was this giant "poster" made up of sheets of paper with fiducials (the dots), each page representing an object (in memory) that could perform some computation. So a few of these sheets of paper, working together, might enable the "identify fiducials" behaviour, by communicating with each other, each of them doing a tiny bit of computation. Some of them might represent some piece of supporting tech, like a sensor (eg camera) or an actuator (eg projector), and act like an API to that thing. One of the things you could do in this setup was point a keyboard object (with its own fiducials) at one of these sheets of paper, and edit the wishes and claims that constitute the computation that that sheet of paper possesses. That means if there's actual """code""" printed on that piece of paper, the code is now out of sync with the real "object" (in memory), but that's fine — you could just wish for a new version of that object to be printed as code on paper, take down the old one, and put up the new one, and now they're back in sync. Then, in 2020, after the pandemic, they needed a way to take Realtalk home with them. So instead of doing all this on a poster, they condensed it down to a binder. Crucially, the pieces of paper don't always need to be visible at all times — the paper (or any object) is just a physical representation of its own behaviour. So you can pull out the paper and let a camera see it if you want to work on that object's behaviour, but when you're not working on it, you can tuck it into a binder. The object still exists. (I think there's some nuance here, since some objects do need to be visible in order for their computation to run, but I think that's, like, a choice of trigger, to put it in eToys speak). Now, with Realtalk 2024, they're trying to get further away from the "textual code on pieces of paper with fiducials" bits, by making it so that you don't need (A) the fiducials to track objects, (B) the pieces of paper — bits of behaviour can be manipulated readily by other kinds of representation, eg drawings — and (C) the textual code, or at least not the sort of textual code they've been using thus far, which is basically some existing language (python, JS, C, etc) with a DSL for wish/claim.

I think Ivan has it mostly correct, except the binder is actually a RealTalk object that contains the other objects within it (that implement the pages). Bret doesn't show it in the making of video for the binder, but I believe there is some process of "moving" an object from a page into a "binder collection object"

Ahhh that makes sense!

(well except for the core reactive loop that implements the wishes/claims system)

re: clause cards (which was a group project, fyi) https://dynamicland.org/archive/2019/Clause_cards https://dynamicland.org/archive/2019/Clause_card_music_sequencer re: working remotely https://dynamicland.org/archive/2021/Supporter_maps https://dynamicland.org/archive/2022/Distant_Page re: “it seems much easier to not make messes in digital spaces” A basic hypothesis of Dynamicland is that we’ve collectively made a big, awful mess in our shared digital space. It’s so easy to sweep complexity under the rug and pretend it doesn’t exist when it’s all neatly packed into the PulseAudio source tree or whatever. Using physical space enforces a discipline of simplicity and understandability. (So it is, in a sense, a “hairshirt”, but we wouldn’t agree that “you don’t need [it] in digital spaces”. Also, it’s a very comfy and pleasant hairshirt. :D)

Thanks @Joshua Horowitz, I didn't mean to imply it was only Omar's work

@Naveen Michaud-Agrawal

I dont think the objection is to small screens, but to the way we do programming today (ie in dead languages with a edit/compile cycle).

But there's a difference between say Glamorous Toolkit and Dynamicland. I imagine a small screen chafes similarly to using a keyboard?

(BTW: I’m just offering the Dynamicland catechism for this. Lots of interesting places to poke & prod into that dogma.)

Similar to how “static types” are not so impactful when you can directly “REPL” your way through the implementation in the living system, is there any summary to why causes (when) + effects (claim / wish) work so well for learning and tweaking existing programs?

@Cole I think it's because you see the results immediately after updating any code

Does folk computer also follow a pattern of cause (like query matching?) and effect? Is that specific language affordance what you are referring to as working better because it is a live system?

Yes, and also using similar nomenclature in the language (wishes and claims). It's built on top of TCL instead of Lua

I also agree that https://dynamicland.org/2024/Is_Realtalk_open_source was one of the very interesting bits of this update. They clearly want to initiate a different computing paradigm and in doing so they are seperating from popular patterns of interactions in the computer world.