Journal — Open source
How this journal uses sources
Dates, source records, AI-assisted work, and the distinction between a project account and a reproducible result.
This journal documents how Digital Sidewalk Lab's research and tools have developed. Its sources include publications, project files, field notes, and working records produced with AI assistance. The entries are retrospective accounts, not recovered diary pages.
Each entry identifies the period of the underlying work separately from its publication and revision dates. The early journal edition appeared in September 2026; a work period such as January 2026 describes the project being discussed, not an earlier publication of the article.
Reading a status label
A proposal describes intended work. An artifact is something produced, such as a specification, dataset, or code package. An evaluated result needs a method, inputs, and an output that supports the stated claim.
The source list identifies what supports each account. Public links let readers inspect published material. An author-held record identifies a private source but does not provide the same public access. Where this edition summarizes a working analysis that has not been rerun, the article attributes its figures to that analysis.
Writing from historical material
A retrospective can explain how a decision looks today, but it should not place that later understanding inside an earlier quotation. This revision uses paraphrase for personal messages whose exact wording has not been checked against the originals. Interview accounts are summarized as presented in the cited research.
AI tools have assisted with specifications, code, analysis, and editorial work. Agreement between tools does not establish a scientific result. A reproduction needs the recorded inputs and method, and a claim remains limited by the evidence supporting it.
Keeping the record useful
Corrections should update the affected statement and its revision history. Source artifacts retain their own dates and versions so a reader can distinguish the original work from a later interpretation.
The journal publishes research accounts and selected artifacts. It does not publish private transcripts, proprietary methods, or confidential project material. The aim is to give readers enough information to understand the work, locate its available evidence, and see what would be required to extend or reproduce it.