Dynamicland FAQ

Is Realtalk open source?

The term “open source” encompasses at least four different meanings, which we need to disentangle:

So the question is actually four questions, and their brief answers are:

Implicit in the questions is the expectation that software be compatible, both technically and socially, with the current culture of computing. Realtalk represents an attempt to grow a very different culture of computing. Principles that we see as intrinsic to the new culture might be seen as violations in the old culture.

Two of the most significant cultural differences might be summarized as “open source is not open to most people”, and “Realtalk is not made of source code”.

Open source is not open to most people.

Source code in a git repo is not open to everyone. It’s open to the select class of people who know what it means to clone a git repo.

In the years that we ran the Dynamicland community space, we observed all sorts of people get their hands on Realtalk — about a thousand visitors of different ages and backgrounds — and the most compelling and forward-thinking ideas came from people who were not at home with git repos. In lieu of sophisticated programming, they added simple programs to interesting materials that interacted with each other in interesting ways, and made interesting use of physical properties and physical space. In wanting to use their bodies and hands, they pushed Realtalk in unexpected directions. In addition, these people were coming to Realtalk with interests other than computing itself, and brought Realtalk to bear on a variety of subject matter.

Git-oriented people, in general, would use pieces of paper as if they were computer screens, typing in monolithic programs to draw images and text. For those whom we had an extended time to work with, it could take weeks of hands-on coaching and immersion in the Dynamicland culture before they started to catch on to the Realtalk spirit. [more]

Most people who sought out Dynamicland after hearing about it online were of this second type. Bringing in non-git people required deliberate ongoing outreach on our part. Most effective was finding an “ambassador” at a non-git-oriented institution such as an art college, science lab, or youth group, who could then bring others.

Realtalk on GitHub, if such a thing were even possible, would open it up, both technically and socially, only to the computing culture that we’re trying to create an alternative to. There would be no opportunity for hands-on coaching or immersion in a community, and most people would use Realtalk in physical isolation (like a personal computer) instead of in a communal setting. Their resulting work would redefine “what Realtalk is”, and would eclipse the delicate and difficult progress that we’re attempting.

Realtalk is just one component of a culture, and downloading source code does not download values, norms, practices, and tacit knowledge. We intend the culture to spread in a manner similar to scientific practices, trades and crafts, martial arts, spoken language, and so on — in-person immersion in a community of practice, teachers teaching teachers. This will take time, and it may appear that Realtalk is “exclusive” during that time. But open-source software is also exclusive, to those who find meaning in source code. And those people already seem well-provided for.

Realtalk is not made of source code.

In Realtalk, a “program” can be any physical situation that can be sensed with sensors and interpreted by another program. A passage of text on a piece of paper can be a program, but so can a hand-drawn diagram, or a poster collage, or an arrangement of objects on the table. Or any combination — a diagram drawn into text, bits of text arranged on a diagram... Or something completely unlike any of these. Realtalk welcomes any representation of computation which can be sensed and interpreted.

These programs all exist in physical space, and are arranged on tables, on walls, on shelves, on posterboards, in binders, in drawers, or hung from the ceiling. Programs see and respond to where other programs are, and arranging spatial relationships among programs is itself a kind of programming. Everything is somewhere, and one can point with their finger to the physical location of every bit of data. Programs can be rooms, or buildings.

This relentless physicality is essential to establishing a communal environment, and to Realtalk’s drastic reduction in complexity. But it also makes it possible to progress beyond the hegemony of purely textual programming.

At present, most of the programs in a Realtalk system are in fact pieces of paper with text on them. In addition to being a research project, Realtalk also has to work, and building working systems out of text is well-understood. This is typical of pioneering technologies — they get off the ground using older well-understood technologies that they will eventually displace. [more]

But unlike text files, pieces of paper with text don’t just carry text. As physical objects, they can also be drawn on, they can host physical tokens, they can be arranged in networks. The sensors see these drawings and arrangements, and it takes very little to interpret them as part of the program. In this way, non-textual programming is never more than one step away.

The most exciting developments in Realtalk have been when people take that step, and represent some part of the system not in source code. When we envision a future Realtalk, far down this path, we imagine every program implemented in a multimodal physical form that combines the best of text, pictures, tangibility, and arrangement, in which the program’s ideas are the most clear and accessible.

A Realtalk system is not, and cannot be, represented as a directory tree of files, the common currency of open source. An attempt to do so would prevent the very progress we’re trying to make.