I think GP was referring to petty vandalism rather than hyperbolic debate, i.e. the tendency of unwatched drying concrete to attract drawings of cartoon penises and bendy swastikas.
There really should be a name for the phenomenon if there isn't already.
> There really should be a name for the phenomenon
I’d say it’s a sort of affordance[0], but I’m sure Don Norman[1] suggested material properties can be described as the “psychology of materials” in his book[2]. He describes how British Rail kept building tougher and tougher glass that was still always being broken, but when it was replaced with plywood the vandalism stopped, even though the wood would be easy to break. Certain materials (and designs) beg to be used (and abused) in certain ways.
The core insight, in my eyes at least, of the Dynamicland project is that a projector + a camera is more interesting for deploying computation in a manner that permeates the physical world than screens. Weiser’s vision [0] could not see beyond the boundaries of the display and unnecessarily constrained itself as a result. But we have cheap megapixel CCDs and projectors now (and the GPU power to get something useful out of “neural network” systems, although I’m not sure whether Dynamicland makes much use of that) and it enables interactions that were unthinkable with a regular two-dimensional display.
And then it turns out that such a system works really well with more constructivist approaches to computer supported learning and collaboration, and the Dynamicland team is doing a very solid job at developing that too.
Interesting, but honestly don't think this is going to be a viable model for education, especially for CS education. The setup looks too cumbersome (and probably unreliable) as of today. And worse, there are just too many hidden parts which are sometimes faulty. This is bad because it creates a leaky abstraction; when the users see some abstraction (e.g. face recognition) reliable but others not, they'll have to work around the system and make up some ad-hoc solution. This might be useful for teaching social sciences (with manipulatable graphs, etc) but I'm afraid that this just confuses students more in certain subjects.
> there are just too many hidden parts which are sometimes faulty
i guess you must live in a different universe than me if this doesn’t apply to “normal” and “viable” computer systems and cs education.
the entirety of your post seems like you are just making stuff up out of thin air simply because it is different than what you know. have you used dynamicland? have you seen presentations on it? have you watched how it works?
if something doesn’t work in dynamicland, you can just go and look at what isn’t working. it’s all there. a major component of bret victor’s work is seeing. watch his talk seeing spaces for some elaboration.
In programming work I often can feel frustrated when I don’t know how to communicate a technical estimate or design nuance to a non-technical manager or stakeholder. If something such as this could be used as a to help express or explore technical planning to non-technical people, I think this could be incredibly useful.
Users here familiar with the game industry likely know about Unreal Engine’s blueprints as a kind of abstraction between C++ code and a more “design-y” format. It has its problems of course, but I think the more we can help express technical problems in a productive way our industry will better manage tech, I think.
I got to play with Dynamicland for a couple of hours last week, thanks to a meetup organised by Steve Krouse and Omar Rizwan. Steve runs the Future Of Coding podcast ( https://futureofcoding.org/ ) and Slack community, and Omar is a researcher at Dynamicland. Last year Omar posted an excellent, detailed breakdown of GeoKit, a map toolkit he created at Dynamicland; it includes much of the technical detail and videos you probably want if you've read this far: https://rsnous.com/posts/notes-from-dynamicland-geokit/
My own feelings about Dynamicland, after a couple of hours there, are mixed but mostly positive:
* Like many who experience it, I don't have a handle on DL's most valuable differentiators as a platform. But that's OK, because it feels like they're not trying to achieve that yet; it's more about just experimenting. We're so unused to computing with this UX model that it's going to take a while to find its best uses.
* Plus, DL is chasing an idea for which the hardware barely exists yet. Ideally, it would be cheaply sensing full 3D in the environment, and projecting higher contrast visuals at all angles.
* Even so, there are some easy, playful ideas they've done that are wonderful. For example, an animator script which lays out three paper-sized outlines to the right of wherever you place the script, and then rapidly cycles the images it captures from those outlines. So, if you want to teach animation, you just draw a cycle across three pieces of paper, put them in the outlines, and the animation shows up instantly on the animator.
* The RealTalk code for the whole DL environment is pasted on a couple of whiteboards on one side of the space. They've voluntarily chosen to constrain themselves in how complex it can be. I really like that.
I really wish they would open source some code and instructions to replicate this. It would benefit many people that will never have the chance to visit Oakland. The closest things I have found are La Tabla (http://tablaviva.org/) and Paper Programs (https://paperprograms.org/).
I see a lot of people focusing on the technology of Dynamicland but less the idea behind it, which is to make computation more of a medium that's continuous with our human world, not something that is trapped in tiny rectangles (hijacking our attention and crippling our bodies) where only people with developed symbolic representation skills can use it. The technology will change and be outdated in many years (like paper and pens have changed over time), but the medium will live on. Maybe my work on Laser Socks provides another example in this space: http://glench.com/LaserSocks/
My point was that it's difficult for an instrument designer to build a piano from scratch, even understanding the idea behind it, without opening one up and seeing how it works first. I understand the Alan Kay-esque hesitation to avoid having a community crystallize around what's perceived by its creators as just a prototype (referencing Kay's unhappiness with Smalltalk's relative lack of innovation over the decades).
I think DynamicLand should abandon their color-dots based encoding system, and use a modified QR Code. QR code is a completely debugged barcode system that doesn't require a color printer. I happen to own a nice big 50lb Canon Color Laser right next to my desk, but in academic environments, black and white lasers are only available, for the simple reason that color lasers cost 4x per page of monochrome printers. My toner cart set is more than the printer cost!
I see one huge problem with Dynamicland. Normal computer programs let you choose how you interact with them. It can be a touchscreen, a keyboard and a screenreader or a screen and voice control / a set of switches. You can write the same code and use the same services using any of those. That's not true about dynamicland. You must be there and actually touch and manipulate it. Whether it's paper and pebbles or some new, futuristic technology, if the main representation sits in the real world, you must be able to manipulate the real world to manipulate the program. That's moving us back about fifty years or so in temrs of accessibility. It's also undoing all the progress the internet has made possible (being able to access and work on anything, anywhere, with anyone).
there is nothing accessible about modern day computers. we can’t see anything that is going on, and it is terribly difficult if not impossible to change things in most cases. almost all of modern computing is a collection of tunnel visioned boxes that can’t communicate with each other.
i just don’t see how you can claim “normal” computer programs let you choose how you interact with them. i don’t see that at all.
i think you miss the point of dynamicland because it is about accessibility. it is about touching. it is about seeing. in dynamicland, there are a multitude of ways to interact with the system.
I haven't experienced it myself, though I've read about it a lot. As far as I see it, the code is actually somewhere in the real world, on a whiteboard, on paper or on anything else. That means that, to change the code, you actually need to be able to touch that paper/whiteboard/whatever. If you have a visual or mobility impairment, you're excluded.
We're confusing two notions of accessibility here. Dynamicland is more accessible for a normal user who wants to tinker under the hood (compared to closed source apps). It's less accessible for individuals who need alternate representations of the digital reality. With Dynamicland, the code is actually stored in the real world and the computer merely interprets it. If you can't manipulate the real world, you can't manipulate the code. When you use a normal computer system, the program is stored inside that system and you can use any input/output device to manipulate it. That includes keyboards, touchscreens, screen readers, switches, voice control or even brain computer interfaces. Using them to manipulate something written on a whiteboard is impossible.
While I agree with the goal of making software and systems more open to the average person, I think solutions like Smalltalk, that don't distinguish between the code itself and it's manifestation and let you open any system and tinker under it's hood, along with better programming literacy taught in schools are the way forward. I agree that smartphone screens with their addictive qualities aren't the best option out there, but I don't think Dynamicland is the solution. Something akeen to it, using AR, or even better, Matrix-like VR when that becomes a thing, would be actually a much better solution. Sadly, we're not there yet.
Brings back memories of tangible interface research around the year 2000, these ideas seemed to have so much potential back then. Unfortunately nobody has come up with a convincing argument as to exactly why we need tangibles, aside from some philosophical musings that fetishize the supposed "richness" of real-world interaction.
The value proposition of tangibles took an especially hard blow when smartphones hit the market — multi-touch direct input achieves much of what tangibles promise to offer (e.g., intuitiveness) while retaining all of the benefits of digital computation (e.g., portability of data, negligible marginal cost of production) that are sacrificed in systems like Dynamicland.
> Unfortunately nobody has come up with a convincing argument as to exactly why we need tangibles, aside from some philosophical musings that fetishize the supposed "richness" of real-world interaction.
i am not sure i understand how appealing to human experience is a fetishization.
> multi-touch direct input achieves much of what tangibles promise to offer (e.g., intuitiveness)
i fully disagree about the intuitiveness of multi-touch inputs which tend to have a hidden gesture language. simply trying every possible permutation of touches and gestures, which is what most people do when presented with a new touch interface, does not represent intuitiveness. buttons, sliders, switches, etc. always yield much more efficient and intuitive interfaces.
also, humans are not digital computers, and we struggle to think like them. humans really like the physical world we're born into, and our senses respond positively to things that trigger them. our senses are dulled when interacting with digital computers. look at the synthesizer world. it's a common fact that software synthesizers are extremely flexible and convenient (in some ways, not so in others) but are monotonous to interact with. people often (almost always?) augment them with tangible external interfaces such as knobs, buttons, and keyboards. people tend to much prefer hardware synths (digital or analog based) that they can touch, feel, smell, see, etc. it creates a tangible environment that is much more enjoyable.
29 comments
[ 3.2 ms ] story [ 88.9 ms ] threadBret already answered similar concerns:
https://twitter.com/worrydream/status/959852576274444288
https://twitter.com/worrydream/status/1036409565389185024
There really should be a name for the phenomenon if there isn't already.
I’d say it’s a sort of affordance[0], but I’m sure Don Norman[1] suggested material properties can be described as the “psychology of materials” in his book[2]. He describes how British Rail kept building tougher and tougher glass that was still always being broken, but when it was replaced with plywood the vandalism stopped, even though the wood would be easy to break. Certain materials (and designs) beg to be used (and abused) in certain ways.
[0] https://en.m.wikipedia.org/wiki/Affordance
[1] https://en.m.wikipedia.org/wiki/Don_Norman
[2] https://en.m.wikipedia.org/wiki/The_Design_of_Everyday_Thing...
Lua!
And then it turns out that such a system works really well with more constructivist approaches to computer supported learning and collaboration, and the Dynamicland team is doing a very solid job at developing that too.
[0] https://web.archive.org/web/20180210180104/http://www.ubiq.c...
i guess you must live in a different universe than me if this doesn’t apply to “normal” and “viable” computer systems and cs education.
the entirety of your post seems like you are just making stuff up out of thin air simply because it is different than what you know. have you used dynamicland? have you seen presentations on it? have you watched how it works?
if something doesn’t work in dynamicland, you can just go and look at what isn’t working. it’s all there. a major component of bret victor’s work is seeing. watch his talk seeing spaces for some elaboration.
Users here familiar with the game industry likely know about Unreal Engine’s blueprints as a kind of abstraction between C++ code and a more “design-y” format. It has its problems of course, but I think the more we can help express technical problems in a productive way our industry will better manage tech, I think.
My own feelings about Dynamicland, after a couple of hours there, are mixed but mostly positive:
* Like many who experience it, I don't have a handle on DL's most valuable differentiators as a platform. But that's OK, because it feels like they're not trying to achieve that yet; it's more about just experimenting. We're so unused to computing with this UX model that it's going to take a while to find its best uses.
* Plus, DL is chasing an idea for which the hardware barely exists yet. Ideally, it would be cheaply sensing full 3D in the environment, and projecting higher contrast visuals at all angles.
* Even so, there are some easy, playful ideas they've done that are wonderful. For example, an animator script which lays out three paper-sized outlines to the right of wherever you place the script, and then rapidly cycles the images it captures from those outlines. So, if you want to teach animation, you just draw a cycle across three pieces of paper, put them in the outlines, and the animation shows up instantly on the animator.
* The RealTalk code for the whole DL environment is pasted on a couple of whiteboards on one side of the space. They've voluntarily chosen to constrain themselves in how complex it can be. I really like that.
I see a lot of people focusing on the technology of Dynamicland but less the idea behind it, which is to make computation more of a medium that's continuous with our human world, not something that is trapped in tiny rectangles (hijacking our attention and crippling our bodies) where only people with developed symbolic representation skills can use it. The technology will change and be outdated in many years (like paper and pens have changed over time), but the medium will live on. Maybe my work on Laser Socks provides another example in this space: http://glench.com/LaserSocks/
i just don’t see how you can claim “normal” computer programs let you choose how you interact with them. i don’t see that at all.
i think you miss the point of dynamicland because it is about accessibility. it is about touching. it is about seeing. in dynamicland, there are a multitude of ways to interact with the system.
We're confusing two notions of accessibility here. Dynamicland is more accessible for a normal user who wants to tinker under the hood (compared to closed source apps). It's less accessible for individuals who need alternate representations of the digital reality. With Dynamicland, the code is actually stored in the real world and the computer merely interprets it. If you can't manipulate the real world, you can't manipulate the code. When you use a normal computer system, the program is stored inside that system and you can use any input/output device to manipulate it. That includes keyboards, touchscreens, screen readers, switches, voice control or even brain computer interfaces. Using them to manipulate something written on a whiteboard is impossible.
While I agree with the goal of making software and systems more open to the average person, I think solutions like Smalltalk, that don't distinguish between the code itself and it's manifestation and let you open any system and tinker under it's hood, along with better programming literacy taught in schools are the way forward. I agree that smartphone screens with their addictive qualities aren't the best option out there, but I don't think Dynamicland is the solution. Something akeen to it, using AR, or even better, Matrix-like VR when that becomes a thing, would be actually a much better solution. Sadly, we're not there yet.
The value proposition of tangibles took an especially hard blow when smartphones hit the market — multi-touch direct input achieves much of what tangibles promise to offer (e.g., intuitiveness) while retaining all of the benefits of digital computation (e.g., portability of data, negligible marginal cost of production) that are sacrificed in systems like Dynamicland.
i am not sure i understand how appealing to human experience is a fetishization.
> multi-touch direct input achieves much of what tangibles promise to offer (e.g., intuitiveness)
i fully disagree about the intuitiveness of multi-touch inputs which tend to have a hidden gesture language. simply trying every possible permutation of touches and gestures, which is what most people do when presented with a new touch interface, does not represent intuitiveness. buttons, sliders, switches, etc. always yield much more efficient and intuitive interfaces.
also, humans are not digital computers, and we struggle to think like them. humans really like the physical world we're born into, and our senses respond positively to things that trigger them. our senses are dulled when interacting with digital computers. look at the synthesizer world. it's a common fact that software synthesizers are extremely flexible and convenient (in some ways, not so in others) but are monotonous to interact with. people often (almost always?) augment them with tangible external interfaces such as knobs, buttons, and keyboards. people tend to much prefer hardware synths (digital or analog based) that they can touch, feel, smell, see, etc. it creates a tangible environment that is much more enjoyable.