back
61 comments
I had the pleasure of visiting Dynamicland and hanging out with Peter Norvig, Nicky Case and many others. There's a few things that stuck out for me:

- The community! Most people I interacted with were pretty inter-disciplinary. Some came from an education heavy background, others from a more programming language background. Some from interface design, others from physical installation or video game background (and everything in between). Everyone had their own degree of skepticism but were excited by the discussions that were happening (a sign of a good research culture!).

- The representations: What excites me the most about Dynamicland is that it's an environment for fostering unique representations of ideas. The cultural forces & ideas (social programming, remixing, visible state / code at all times, 3d environment, etc) baked into the space encourage the experimentation & creation of new ways of representing complex ideas. Some scattered examples here - https://twitter.com/dynamicland1?lang=en

- Removal of artificial barriers: When programming on a computer these days, there's so many barriers to doing simple things. The amount of code that exists to do virtual actions in a virtual world is MASSIVE, and acts as a huge barrier. This is something Bret talks about at the end of his interview on Postlight (https://postlight.com/trackchanges/podcast/computing-is-ever...). Because code / programs in Dynamicland are embodied in the room, you don't need code to move a dialog box or a slider around. You just moved it yourself. This is super powerful and it means your code can focus on the actual computation itself, not on virtual UI movement. Eventually, Dynamicland will have robotics to automate moving of objects, but this already is a great start.

- Moving around: Moving around, even if it's just around a table, is AWESOME. We're so used to sitting (or standing) at a desk and staying still that we don't get to take advantage of embodied intelligence. We think in SO many different ways, and our "thinking" doesn't just happen in the head. Our arms, legs, stomach, and feet all contribute to the thinking process. Combine that with multi-sensory representations that live across all of these channels and you can explore a thought & idea space SO quickly and uniquely. This is so hard to describe and under-rated. This link (http://worrydream.com/ABriefRantOnTheFutureOfInteractionDesi...) and this talk (https://www.youtube.com/watch?v=agOdP2Bmieg) attempt to do these ideas justice but it's quite hard to transfer this context!

Is it possible to visit Dynamicland as a random interested person? I'm on their mailing list and couldn't make the open house last spring - maybe they'll have another one at some point.
I made a small donation to Dynamicland and then requested to come by as part of that.
When I visited Dynamicland in January, I was building on a spreadsheet-like program next to my friend Omar who was building a map-based interface. Just by virtue of sitting next to each other, we were able to keep apprised of what the other was up to. At some point Omar needed a way to input a number to control the zoom of his map. My spreadsheet program had plenty of numbers, and ways to manipulate numbers, so we slid over one of my pieces of paper and it immediately worked to zoom into his maps. Omar decided that he’d prefer a slider-based number-input, and after it was built, he slid it over to my side of the table, and we used a multiplication operator I had built to expand the slider’s range. Again, it just worked. At Dynamicland you get composability and interoperability “for free.”

I jumped around a skimmed a bit, so maybe I missed it, but are they prrogramming for the room itself only, or in a framework/library that allows for this, or just in a similar language? Is it all just Javascript and a well defined hierarchy of objects that can be manipulated?

How do two people, working on two separate programs, share implemented functions "for free"? The sliding across the desk portion is irrelevant (well, it's cool, but for this specific question it's no different than sharing a gist link I think), what I'm wondering is the details.

Omar's article has all those details for you https://rsnous.com/posts/notes-from-dynamicland-geokit/
Ah, thanks. For those following along, the details (from what I've found so far) seem to indicate a language they've developed at Dynamiclab for this, Realtalk is used and embeds the idea of the whole platform (papers that can be both sources and targets of information/code) into itself, and how they interact with each other.

So, yes it's a shared language, and yes, it's a shared framework (the language embeds the framework concepts as first class entities in how it functions. That makes sense, and it also explains how some output display function (which would be fairly generic in this context) would be easily shareable, and by the nature of the platform, also immediately shareable if done in a certain way.

It is similar in concept to how Javascript ona webpage has access to the dom, and the bindings for the dom can be expected, so you can write something that transforms a <table> in some way, and expect it to function similarly is other tables are provided. But it might be even more accurate to say it's like CSS, where CSS and the DOM are so closely linked that (at least from the perspective of CSS, if not HTML elements in specific) there is no interop layer, CSS is meant to apply to an HTML document, so it's designed with that in mind and the interop layers are for the most part nonexistent.

So, in that respect Realtalk is sort of like CSS (with more programmability, I think) for this environment of sheets of paper that support input and output. Very cool.

At least when I visited (which was about 6 months ago I believe?), it was a fairly basic set of extensions on top of Lua. As in, they've added a couple keywords that I believe they just RegExp for (its pretty simple to confuse the language parser) - the language itself seems to not be the focus (yet?) of the project.

Again, I only used it once (and I made a DynamicLand compiler in DynamicLand! Video here: https://twitter.com/i/web/status/963497112284512256 ), but if I recall correctly, the main way to "share functions", was to actually just load someone's code and copy it. Partially because the language/editing seems very new still, and perhaps partially to encourage this sort of "real world sharing" vs. virtual sharing. I recall just "forking" files a lot and making new copies of things (kind of like "Old" folders I used to make before version control).

"Light is not a device we charge up and carry around in our pockets. Imagine how much dimmer that world would be: people carrying flashlights, shooting small cones of light wherever they go. It would be a small, lonely, personal world, a world where we only get to see one thing at a time, a world where one of your hands is always full with an electronic gadget." Nicely said
I think it's more accurate to say light is not only a device we charge up and carry around in our pockets. I do in fact always carry a flashlight in my pocket and use it fairly regularly and many people use their smartphone as a makeshift flashlight from time to time even if they don't carry a dedicated flashlight. I imagine the extended analogy would likely hold: if Dynamicland type technology becomes common I expect it would co-exist with more traditional computing as well as other newer tech like VR/AR.
Sounds terrific. However, it feels like a lot of what they have achieved is marketing existing ideas that were already great, with improved packaging, but in a way that ultimately isn't practical.

Because truly plug-and-play components are an awesome idea. However they are not a new idea at all. As far as I can tell the real reason they are not more popular is because us programmers are worried we will be accused of being users if we use tools that are interactive. See Visual Basic 6 (which was an amazing system). Of course programmers do not realize this psychological issue exists and will not admit it is a possibility.

But in the context of Bret Victor's lab with the projectors and little pieces of paper, that puts plug and play components in a different category that makes them more palatable.

Also, projector based AR is much more convenient for people to demo than HMDs, but also not very practical for widespread deployment. But AR is amazing in general and this allows them a friction free way to demo its power with components.

But I think that multiplayer real-time collaborative interactive component based AR programming should definitely be a more common thing. It combines the advantages of all of those techniques. I also think though that all of those things are useful in lesser combinations. I believe a big part of the reason at least some of those things are not used more or more effectively is cultural or psychological rather than technical or practical.

The thing is the more you take advantage of components and interactivity the less you use the complex colored encoded text. That means programmers can spend a significant amount of time snapping together components and configuring them. Unfortunately programmers are not able to recognize this as programming. We have a feeling the more we do it the more we may lose our special incanter status and be classified as users.

> us programmers are worried we will be accused of being users if we use tools that are interactive. [...] Of course programmers do not realize this psychological issue exists and will not admit it is a possibility.

It would be nice if you explained what you even mean by this instead of just claiming that programmers who don't agree are in denial.

What programmer is embarrassed to also be a user? And programmers use interactive tools all time. REPLs, debuggers, editors, GUI design tools, and live-reloading web servers all qualify in my book.

Also keep in mind that when you stack abstractions on top of each other, the whole stack gets more brittle. So if you want me to use your better idea, it shouldn't abstract over the thing you're trying to replace.

To give an example: if you want to replace hierarchical filesystems you shouldn't build a thing that hides them from me, you should build a thing that doesn't use them at all. (This isn't a hard rule, sometimes exceptions are worth the trade-off.)

Understandably, this means changing things is a LOT more work than you're probably prepared for. Don't assume that we're only stuck in the past for bad reasons.

if you need any example of this type of behavior just look how "normal programmers" view labview programmers. most people completely snub their nose at labview "users" for not being real programmers and engineers (don't tell me they don't because i experience this constantly). in addition, labview's capability is often misunderstood by people who hide behind a wall of self-imposed bias. many times people are actually unaware of what is possible with labview simply because they don't care to know because it's not the real, hardcore, in the weeds type of stuff they were taught in school. meanwhile, people using labview to program windows, real-time high-performance, real-time embedded linux, and fpga systems (including cameras, sdr, etc.), all with the same language, are able to accomplish quite a lot. i fully concede that labview has many downfalls, but my criticism is based upon actual use of the language and tools and not via some uninformed bias. because of that, my criticisms don't necessarily overlap with those of the uninformed. and this is all relevant because when i tell people i use labview, you can just see any interest in my programming work leave their face, as i've become a simple tool user in their eyes.

this is a specific example regarding labview, but these types of biases exist everywhere, where engineers blind themselves to what is possible by the opacity of what was possible.

I wouldn't want to write a long letter on a piece of paper never mind software, it's not because I'm addicted to a fake status symbol it's because "reality" in the form of solid objects has limitations that software doesn't eg "a delete option". And the rest of vim. At Stone point you have to encode complex rules, and having a slider framework linked to moving a rock in front of a camera isn't going to help you write a compiler.
Can I build a Dynamicland clone in my area? That'd be excellent. For all the press this thing gets (from time to time), I still see neither designs, plans, nor code that I can use. Is this because Bret Victor shares ideas and expects that if the world finds them valuable, they'll implement their own? Or is there another motivation to hold this close?

Maybe someone has to build an Open competitor so that we can address problems mentioned by other commenters. Maybe I'll do that here in my hometown and share it with everyone.

Jan Paul beat you to it https://news.ycombinator.com/item?id=16211516
Shut Dynamicland down and focus all Bretts and the rest of the teams energy on building these collab interaction ideas into AR.

Dynamicland has 2 components - the paper and projector idea which is bad, and the interactive OS and systems which are fantastic. Replace the paper and projector with AR glasses and you have a winner.

AR allows exactly what you want - real time dynamic collab 3d interactions. Paper does not. It's 2D, flat, limited, with occlusion. You've just taken 2D screens and reversed the light source.

Dynamicland is a sunk cost. Drop it fast. Move to AR.

I was thinking the same thing. Does this technology offers anything more than AR? It seems the answer is no, unless you really don't want any glasses on your face.
Everything old is new again. https://www.youtube.com/watch?v=laApNiNpnvI (Xerox EuroParc 1991)
It's funny you say that. This is something Alan Kay talks about at length. For example, watch the first 20 mins of this talk - https://www.youtube.com/watch?v=FvmTSpJU-Xc

Essentially, he asserts that computing (or computer science even) isn't a real field (like biology or physics). Most people can't name the early computing pioneers whose work they build on (not true in Physics or Biology, we celebrate the pioneers) nor are they familiar with the work that was done in the Xerox PARC days.

It's gotten so bad that a lot of computer science research just assumes that what we have now is what's going to remain forever. In the 70's, all kinds of interesting ideas and experiments were tried.

Hardware implemented VM's, direct manipulation programming, OS-free computing environments, highly re-configurable computing a la FPGA's, and hundreds more.

Nowadays, we think that creating custom ASIC to run machine learning algorithms quicker is innovative and novel.

All this to say that many of the ideas in Dynamicland aren't new. They're rooted in ideas that are decades old. If you look at Bret's papers he uses as references frequently, you'll notice how many of them are over 10 years old - http://worrydream.com/refs/

my jaw literally dropped at 3:12. https://youtu.be/laApNiNpnvI?t=192 I still can't figure out how they managed to get the camera quality/resolution in '91. I'm guessing that it's just an image rather than conversion to text if it's a paragraph. [edit] what happens is copy and paste from a book on the desk to the computer clipboard. where the opposite corners of the physical screenshot are defined by two fingers creating a rectangle screenshot area
This seems to be the next Xerox Parc! I tried to find some explanatory Videos but wasn’t successful. Has anyone some links to this highly intriguing experiment?
Physical computing sounds like an incredibly limiting environment, like using FrontPage rather than HTML, CSS and JavaScript. How do you do anything abstract on a platform like this? How do you i integrate any two arbitrary pieces of software? How do you apply basic programming practices such as functional (as opposed to copy/paste) reuse? Or do you end up modifying code the 95% of the time the abstraction hinders rather than helps?

I really hope this isn't how we teach programming to the next generation, because it will severely limit their understanding of what makes programming great.

I would look at this at opening up a new additional way of computing, not limiting how computing works in general.
I wish I could work with people who are as interdisciplinary as this team. As an interdisciplinarian myself (psychology, busses, computer science and game-design) I need a place to belong and shine creatively.
The Sony Xperia Touch ought to be able to do this. That's a nice projector-based touch system.

Possible application: interactive restaurant menus. Some places have tried handing out tablets, but you have to have someone hand them out and retrieve them. The system has a camera, so have it recognize open space on the table and people in seats. Present people with menus projected on unused table space. Plus you can offer customers games while they wait.

(Inevitably, someone will put in ads.)

This looks amazing. Such a refreshingly different take on things, and such a beautiful -- and realized! -- positive vision for how we might interact with technology and each other. Brilliant.
How does one get a tour of Dynamicland? Is it open to visitors and/or is there some way to get invited?
I’m not understanding how this works. What actions let you create a new table in Postgres or drop an index? Are there ways to install new software like pip?
As far as I understand Dynamicland is not a general purpose OS that tries to do everything. If you want to manage PostgreSQL, the shell is a pretty good way to do that.

Dynamicland looks like a real world IDE, where a lot of the interface consists of physical objects.

It's more like a Mathematica notebook than a bash shell.

Imagine Microsoft Bob but in real life. If you need assistance, just ask Clippy. He's projected wherever you go.
> At Dynamicland computing is social like cooking is social. It’s also physical like cooking is physical. You’re not seated in front of a single-person screen, but walking around an open space, using a range of tools.

Ya lost me. I hate cooking specifically because of the physicality, of the different tools, etc. If I liked cooking and all that shit, I'd be a chef.

>If I liked cooking and all that shit, I'd be a chef.

It’d a pretty dull existence if the only things you liked to do could be determined solely by your current job

I love cooking and I love programming (as a career and a hobby) and I love doing many other physical things. That's my anecdata for what it's worth.
This works great when a small, gifted and creative community has tight controls on the process.

But how will it work when non-creative people take the technology and make it bland — and infuse it with ads, because that is the only viable business model that tech seems to be able to come up with.

This is always a risk with inventing something new. Their FAQ hints at how they plan to mitigate some of the negative effects - https://dynamicland.org/faq/

They want to mirror the Carnegie library model as much as possible. They're also non-profit and not interested in building a product that can be purchased (instead they think of this technology as infrastructure available to almost everyone, like the internet or running water).

Thanks for an excellent article Steve, I thought it captured the spirit of the place quite well.
Does dynamicland have any kind of permissions / access system, also with regards to pieces of the room, or the whole room, or is root access to the system also controlled physically, like a conference room projector?
no, on purpose. I single person can crash the entire system (it reboots quickly). A single person can spam the entire facility. This is, at least the moment, intentional.
i don't see the big difference other than just virtual interacting, other than lots of clutter; and as miki123211 said..accessibility seems like it would be quite an issue which is very important to me
Still not open source. Or "beyond" open source.
Really? Red thin text on white background? I'm sorry, I will not read it. Is this the purpose of this site? What's the sense in writing something normal people cannot read?

And yet it gets to the 1st page on HN, no one complains, and my comment will get huuuge number of negative points, I know, as usually when someone posts a comment about this kind of bad design of a page/author people worship.

I liked the red text. It's a very well designed layout, and the color choice - while arguably questionable - gives it personality. Not every site needs to look the same.

The fact that it has a lot of upvotes shows that aparently many people don't have issues with the presentation.

Anyway, my browser has a "reader" mode that lets me make any website readable with a single click, so I've stopped complaining about hard-to-read aticles. Better to focus on the substance of the article...

... so .... virtual legos?
It's nice and all, but this is a nightmare for accessibility. Not everyone can read paper or even move around. With this design, coding around those limitations is not easy. In the current model, everything could be done 100% accessibly. Why it isn't is a completely different problem. When everything is inherently physical, though, accessibility is not just about the code, it involves much more parts, such as automated devices to move things around. Considering how expensive devices made for accessibility are (a braille display is usually well over $1000), that would be a huge setback. I think the beautiful thing about computers and technology is that you're not tied to any physical medium and that information can flow freely between devices that can represent it in different ways. Now, there can be ten programmers working on a codebase, where one person uses a normal Windows box with a lot of GUI editors, another one uses Linux in text mode with vim, a third one has a Mac Mini with a screen reader and no monitor attached (my also blind friend programs like that), and the fourth uses switches, eye movement sensors or voice-recognition technology because they're physically disabled (see https://youtu.be/YRyYIIFKsdU?t=501). In other words, when everything is purely digital, you can consume and manipulate information in any form you like and in any form the computer can work with. It can be on a smartphone screen and with your finger, on a TV with a remote, through a voice-controlled speaker or a laptop with a keyboard, the choice is yours. The information is purely digital and not tied to a particular medium. However, in Dynamicland, the information is mostly physical and it's the computer that needs to process it, not the other way around. That makes consuming the information through a different medium than it was originally presented with very difficult. It would be possible for a computer to describe a Dynamicland environment for the blind, but it wouldn't be easy for a blind person to write the code on Paper. In a normal computer environment, such a person can just use a different tool (AKA a screen reader instead of a display) and work on the level of their sighted peers. However, a dYnamicland environment makes such things impossible.

I recognize the limited Value of Dynamicland as an exercise environment, maybe like Scratch. Even in that form, special considerations would need to be put so that disabled people have an alternative way of doing the same exercises.

Dynamicland makes a lot of other things, not related to accessibility, harder too. For example, how do you deal with larger codebases or distributed teams? How do you quickly collaborate on one piece of paper while being in two different places? Very simple to do for a Google doc. Almost impossible to do here.

The idea is kind of interesting but it would be only worth it in a pre-internet age. In 2018, people expect realtime collaboration with people in different parts of the world and with Dynamicland, that's not going to happen.

I wrote to Dynamicland's email address about this months ago, and even re-sent the email, but never got an answer.

I suppose it's inevitable that if this vision of the future takes off, we'll be second-class participants. But I guess it's too much to ask the majority to not work in a way that would be great for them, just because it's a new challenge for a few of us. We'll figure it out.

Similar accessibility problems already exist with the current interface of keyboard, mouse, and screen. If you have a physical disability making it impossible for you to type, use a mouse, or, are blind or colorblind, you require specialized software or hardware.

There is no reason why Dynamiclands can't similarly have specialized hardware and software for people with various disabilities.

Plus, a more natural/human interface might be more _accessible_.