Do we have the same idea of what an Open App Ecosystem is?
Bob Haugen
Public
Seen by 391
I see two or maybe three different ideas floating around:
1. Collect a bunch of existing open-source software and add a single-signon app and maybe a dashboard and some navigation. This is good and useful. Sorta what DigLife is doing (but see also #2).
2. The next stage is what DigLife adds to their collection of apps. Notably, in what I saw, a bot that does some integration of the apps into somewhat more of an ecosystem, and possibly some data-sharing among the separate apps.
3. Yet another stage is radical decentralization, which will require creating some new apps, because what is needed may not exist.
I think Holochain is an example of this, which they describe as agent-centric instead of data-centric, as is the proposal that @lynnfoster recently posted called Structuring the OAE around agents. This is also similar to the Indie Web, Solid, and SSB. In all of those visions, you own your own data and your own identity and you may interact with other agents, some of which may be individuals and some of which may be groups that you voluntarily participate in. Agents in this stage could also use different existing apps, but their main interactions would be with other agents, not with other apps.
I don't think any of those directions are bad. But they are different.
Which of those visions are you working toward? Or are you working toward something else? Or do you think I mis-stated the question?
Greg Cassel Tue 19 Dec 2017 3:12PM
I'm definitely working on something more like #3 than the other 2. I don't call my models agent-centric per se but I could. They're about relationships: between data items, between items and agents, and between agents. (They're focused on peer-to-peer relationships, but they enable the development of social hierarchies/ levels of organization in clearly-defined contexts.)
When people say "agent-centric", I guess they're trying to restore true relationships between agents and data items, instead of pretending that data can exist in a vacuum. I support that. Certainly I support each agent having the opportunity to create digital identities which they control, and building personal networks (with both private and public components) around those identities.
If my models are deeply agent-centric, they're emphatically community-centric on top of that. They're about identifying and building communities of peers-- potentially even global communities with practically no entry barriers. They're about enabling and facilitating community functions as simply and flexibly as possible, without requiring a wide range of mostly redundant apps to perform basic signal routing, processing and storage actions.
Frankly I'm pretty pessimistic about most efforts to create compatible ecosystems between specific existing apps (and their teams). However, I think that concepts #1 and #2 in this post can be valuable if they're oriented toward creating a truly open and inclusive ecosystem, instead of simply reinforcing mutual benefits between existing development teams.
Bob Haugen Wed 20 Dec 2017 12:39PM
Greg, a couple of questions and comments:
What does "community" mean to you? A geographical community? Or some kind of organization?
If organization, does the org have agency? Can its members agree together make agreements with other organizations, or to engage in economic interactions (like maybe set up a software hosting environment or a cooperative kitchen or farm)?
If so, it is an agent in the agent-centric model we have in mind.
Re compatible ecosystems between existing apps: DigLife appears to be doing so, and we have an experiment in FairCoop where a bridging app is taking information from two other existing apps (GitLab and NRP) using the ValueFlows vocabulary and using the combined information for income distribution. We'll see how it works, but it seems to be possible.
Greg Cassel Wed 20 Dec 2017 6:02PM
@bobhaugen , I personally define community in a painstakingly generic way: A community is a network of agents who all share a specific relationship with a specific resource.
That definition can apply to all geographic communities and all organizations.
I also define agent and group agent in painstakingly generic ways. "A group agent is an agent which represents two or more personal agents within an official context." -- I believe that this is entirely compatible with the agent-centric model you have in mind.
Draft Wed 20 Dec 2017 6:45PM
@bobhaugen is a ressource an agent ? Is an organization an agent ?
Bob Haugen Wed 20 Dec 2017 6:49PM
is a ressource an agent ?
In general, no. Gotta have agency.
Also, I don't use the term "human resources" except to point out that humans are not resources unless they are slaves. Humans can provide resources (their work, their ideas, etc.) but they are not resources.
Draft Wed 20 Dec 2017 7:04PM
Totally agree, human are not ressources. What about organizations, are they able to be an agent ?
Bob Haugen Wed 20 Dec 2017 7:12PM
What about organizations, are they able to be an agent ?
Yes, with some complications. They need actual humans to act on their agency. And re @gregorycassel , the humans in the organization need to agree on what the org is doing on their behalf.
But organizations do make agreements with other organizations and engage in economic interactions with other orgs and people, all of which are examples of agency.
Draft Wed 20 Dec 2017 7:20PM
And an organization could also not be an agent ? Or it has to be an agent ?
Bob Haugen Wed 20 Dec 2017 7:30PM
an organization could also not be an agent ?
Yes. All up to the org.
Draft · Wed 20 Dec 2017 6:46PM
I understand.