Victor's Dynamicland vs the Object Network
Techie article! Scoring Dynamicland against my architectural vision
Here I’m going to compare Bret Victor’s Dynamicland against The Inversion - my architectural vision behind the Object Network - so please do read that first. I don’t go into lots of detail about my vision or its justification or why it’s better, this is just about how there’s nothing out there that does what *I* want.
Bret Victor’s Dynamicland
Dynamicland is very specifically local-first: you have to be there, basically. It’s so physical that it has paper instead of screens. Having not visited this far-off Land myself, and as it’s closed-source, I can’t imagine I’ll say the Right Things about it, but I’ll give it my own interpretation anyway, with up-front apologies!
Each paper object is a self-contained meaningful thing that is “live” due to having anything you like projected onto it and following it around. Objects are identified via coloured blob ID codes, which are probably not globally-unique! Objects are directly physically manipulated and can be pushed around to allow interaction with other objects, so there is probably some concept of a link, where an object has the blob ID of another.
Objects don’t actually need to be paper: you can drop in any physical object that can be identified via its coloured blobs. The object types are currently more science lab and IoT than office, but that probably isn’t fundamental.
Interaction and behaviour seem to be “internal” to objects but based on a single local state-assertion database, driven by script fragments that do give a spreadsheet data-dependency feel. So green for Declarative.
It has LAN networking with a peer-to-peer protocol (I think), so earning some Decentralised points, but apparently not WAN, fitting with the hyper-local-first philosophy. [Update: they can have some level of inter-room interaction as seen in this page from the new website]
It’s described as an “OS”, which seems fair enough. I’ll give it amber for 3D because it works in the real world with 3D objects, but not green because that’s obviously not what I meant!
|Deconstructed
| |Declarative
| | |Decentralised
| | | |OS
| | | | |3D
|8 |8 |8 |4 |2 |
--+--+--+--+--+--+
DL|🟢|🟢|🟠|🟢|🟠| 25
--+--+--+--+--+--+Differences to the Object Network
Dynamicland is missing the Metaverse! It could easily be ported directly to an Augmented and Virtual Reality world over the internet (with the same peer protocol it uses over the LAN). Then you would need global identifiers for each object, based on the coloured blobs, and could tie them together with such links, instead of using the more ad-hoc adjacency. Object types would need to be more standardised if you did this.
But of course that would clash with that local, in-person tangibility that is core to the vision and philosophy; Victor rolls his eyes as I undermine the whole point of his project! Those two blobs for Decentralised and 3D being amber in my ratings are what make Dynamicland what it is. Being a Meta-verse would undermine its Real-verse-ness.
But I may as well continue: how about we replace Realtalk’s partly Imperative and scripted programming model with a fully Declarative approach. One that has both default code per standard object type plus also allowing the creation of general rule objects for the internal animation, interactive behaviour and data dependencies between these linked objects.
Dynamicland comes very close to the Object Network, but OnexOS is unlikely to use paper in rooms instead of 2D panels in 3D, or to give up using the internet or AR!
Hit the subscribe button to learn more, and feel free to leave a comment or question!