title: Beyond identity management - rethinking the purpose of personas
tags: holochain, p2p, happ
Personas could be interpolated as lenses through which we watch the world. Such a multi-lens paradigm is enabled by managing identities at the « hApp browser » (e.g. holoscape) level and will allow to create multi-hApp feature-rich containers for each persona that can be further enriched with agent-level AI through local APIs for a transformative and integral Web3 experience.
> [name=pospi] Phew, that's a lot! Can it split into 3 sentences? (;
> At a technical level I am curious what this really means- are we talking about "identities" as key management, aka deriving separate agent keys to switch agent contexts in Holoscape? Just checking my understanding...
## Please bear with it... and comment
This is an attempt to extract from my mind something on which I haven't put to words before. It feels like it's been a key part of what has drawn me to the Holochain world, so I hope to convey the meaning of the thinking clearly. However it might take quite some feedback and re-drafting to get this work to a point where any proposal for the Holochain community or core team can emerge. Please bear with it, and ideally comment inline along your reading.
> [name=nicolas] This is an inline comment example. Scatter them all around :)
## Beyond identity management
> [name=nicolas] I use the words « persona » and « profile » mostly interchangeably here, but we may need to build clearer distinctions as we go.
Looking at the [Personas and Profiles document](https://hackmd.io/pcDkiCJoQH-z6s_VS4LNRg), I understand the Identity Manager was created to « give [users] back agency over [their] data ». It allows a user to create different « personas » to help them separate personal data they'd use to interact in different communities / hApps.
Appart from offering sovereinty over your personal data, this allows you to save time by reusing the same information over and over in multiple places (assuming all the hApps are using the same naming convention for writing their [Profile Spec](https://hackmd.io/pcDkiCJoQH-z6s_VS4LNRg#Profile-Spec)).
This is great stuff, but I think the benefits of having personas should not stop there. Creating personas (or several profiles per persona) enables a separation of concerns that could have many uses beyond the simple scope of identity management. Pardon me the implied techinical stretch (which I need your help to understand), but imagine:
- if Holoscape (or other hApp browsers) had a identity selection mechanism to browse the Web **as profile X** and have all hApps automatically configured for the data and preferences of that profile.
- e.g. You often do personal trips and business trips and create personas for both.
- When you select the personal persona:
- your map hApp highlights possible hikes you can take
- your events hApp has high priority push notifications and lets you know of all the events you'd love doing
- your getting-together/dating hApp has high priority push notifications for helping you finding the people you'd have fun with
- you're automatically sharing your location with selected friends and familly
- When you select the business persona:
- your map hApp shows the restaurants and cafe where you like to meet people for work
- your events hApp shows events you need to attend for your work (meetings, upcoming appointments etc...)
- your getting-together/dating hApp is configured to only notify you of people that would make sense to meet to further your professional goals & unfold opportunities.
> [name=nicolas] @guillemcordoba maybe your understanding of the tech can help assess whether integrating identity management within holoscape and switching DHTs based on which profile is selected would make sense at all?
- if all developers of hApp would think of what information should be accessible by a agent-level local API so that agent-data from different hApps can be aggregated. You could then have things like:
- a mixed activity feed (e.g. last push notifications, subscrpitions, content I follow, events etc...) from all hApps that are bound to a specific profile.
- visual representations of the social networks of a specific persona/profile aggregated over several hApps, seeing what are my relationship to them (e.g. who you are following on tweeter, who's your friend on Faithbook, who are the authors of this persona's favorite podcasts...)
- local AI agents (call these Virtual Agents?) extrapolating a multi-DNA pool of data (collected via mentioned local APIs) applied against a set of desired goals to help you sieze the right opportunities to meet these goals (e.g. new people to follow on XApp, an event to attend, a local farmer to visit, a new personal management tool to try out...). Think of complex data-mining algorithms put to the service of your genuine needs instead of profit-driven advertizing.
> [name=nicolas] @pospi (and others) just some more food for thought on this last point. How about virtual agents (VAs) that would make use of a [MetaIntegral](https://www.metaintegral.com/) manifest to define specific intended impacts, then help you strive to make these impacts a reality and finally give you visualisations helping you to make sense and tell stories of that impact... Also, if we collect and classify Integral Data points in first-person, second-person and third-person categories could HoloREA help us make sense of that data to help the VAs target the desired impact? Or maybe it's more a case to be used for reporting?
> [name=pospi] Sure, that sounds reasonable to me. I actually started a post about this same topic at https://forum.holochain.org/t/implementing-cross-app-agent-centricism/2145 if I understand you correctly?
## Wrapping up
I feel Holochain has the potential of redefining the Web experience. In that perspective, would thinking in terms of « what are the applications we should create » be short-sighted? Can we create novel components/layers of technology (such as Virtual Agents) that assist agents individually by helping them being the impact they want to see in the world, or technologies that help interface DHTs and uncompatible DNAs so that cross-communities synergies can appear?
Please tell me how.
> [name=pospi] We need to get the indexing / filtering stuff sorted, because we need [search features like these](https://github.com/valueflows/vf-graphql/issues/75?email_token=AADK2JEOMEWNNE72I5UKTK3RGUMXFA5CNFSM4LCDRQ4KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEOHZF6Q#issuecomment-598005513) in order to do that kind of analysis. I have been thinking about an iterator API for efficient link traversals which I think could be a nice start- you get a lot of the performance for free with those traits.