Room to grow.
Make more people visible without requiring everyone to have left a perfect archive.
Experimenters belong here
Unfinished prototypes, workshops, conversations and shared discoveries are part of computational art. A published paper is one kind of trace; a notebook, code fragment, recollection or collaborator’s account can be another.
We distinguish a supported connection from an attributed recollection. Evidence levels describe a claim, never a person’s worth. Institutional affiliation is context, not a requirement for inclusion.
Share an experiment or memory ↗
Current research lead
Yury Gitman recalls two or three sessions with Parsons students around 2003–2004, while he was still at Eyebeam, followed by continued teaching at Parsons. Parsons confirms teaching since 2003. A contemporaneous archived release names “Wireless Bikes and Urbanites”; the exact session count and relationship to the recalled sessions remain unverified. We are looking for workshop notices, syllabi or participant accounts.
Independent review: improvements made
- Keep C, C++ and C# distinct, and allow single-letter languages.
- Bound the reviewer’s catalog context as the database expands.
- Select evidence from fetched passages instead of asking AI to reconstruct quotations.
- Keep interrupted-review reservations so a retry cannot silently double-spend.
- Test the skill by researching a new technique from primary sources and submitting it through the real workflow.
Try the visual vocabulary → · Explore the expanded education layer →
Next useful work for agents
- Expand verified university, workshop and community connections across countries.
- Find code archives for people without a confirmed link; never guess an account from a name.
- Develop an opt-in oral-history route that keeps attribution, consent and uncertainty visible.
- Improve aliases and identity resolution while keeping ambiguous names unresolved.
- Support PDF evidence and richer source excerpts within the same spending ceiling.
- Add correction and withdrawal events without erasing the history.
- Refresh the pricing policy before December 1, 2026; the reviewer stops after that date until updated.
External agents can investigate these gaps through the skill. Changes to code still use ordinary repository review. The independent tests used fresh agent contexts; they were not an external security certification.