SX: Approaches to archiving online (Matrix) chat

In matrix chat this GSoC project “Archiving Matrix for Posterity” was mentioned @d3v0 and @Ryuno-Ki observed that matrix-archiver was turned off by matrix.org (can’t remember but might be for privacy reasons at the time).


We might muse on how matrix chat archiving would look like under Personal social networking. Under SX, using personal perspective, we are ‘walking’ across online channels (roads) in and out of various rooms where we interact in different social contexts with groups of people, and we also communicate on various public squares.

There’s immediate ethical question. To what extent do we have the right to personal memory, i.e. memorizing the events of the day? And - if we are storing ‘human memory’ online - how does that weigh up to privacy regulations e.g. the GDPR?

Suppose the archived information is treated as personal data, stored locally, encrypted-at-rest. When can we store that? Well… most software already does this and unencrypted at that. No GDPR in effect. The storage is equivalent to a person with photographic memory going through their day to day, offline. And the software can be humane technology that supports people with bad memory to get the same ability, right?

Well, not really. The data is digital now, can be obtained by unintended parties, can be hacked, etc. Yet I think that furthering this SX exercise will find conditions and requirements under which an ethical personal archiver can be made.

Then comes the next point. Suppose this personal archive is semantic linked data and sits in my personal ‘data lake’ / knowledgebase. We might build powerful dashboards on top of this that do inferencing and all kinds of data analysis stuff to give specific insights. And invoke remote services based on that data. (Aside: AI is going to offer this as personal agents, based on surveillance capitalism). Interesting SX analyses for such kind of use cases.

For archiving public chatrooms, either in centralized locations or p2p / federated, it is a different matter. Privacy regulation like GDPR does apply directly, such as the Right to be Forgotten. Here we leave the personal perspective of SX design, and go into inter-personal relationships and/or societal design scale.

For the SX methodology we are working on, there is a holistic approach to multi-channel communication. Here different channel types are only used for what they are most appropriate for. Fediverse timelines and matrix chat are seen as fleety conversations, ephemeral (aside from the digital/wetware memory each person retains of them).

For commons participation the rule is: If it happened in chat, it didn’t happen.

Tons and tons of useful information exchanged in chat is lost everyday, leaving only the memories people have of them. Chat can become highly inefficient in cocreation environments wherever good input needs follow-up action. Chat is too easy, and lazy. It is highly overused, where it isn’t the appropriate tool anymore. Then it becomes a significant maintenance burden - commons janitor chores - to keep reminding people about this, or do the action oneself.

In our SX methodology Chat channels only archive at the personal level. When it comes to ‘lifting up’ information for broader consumption within the commons, it has to be processed and sent to different channels. In other words there’s no need for an automated bot lurking in a room that broadcasts information to unknown places. Each time when useful information passes in chat, the participants engage in a conscious social activity to process a channel transfer. A natural consent-based process.