The paper trail: how the records were pried loose
Sources: the City’s response letters are in the library under “Freedom-of-information correspondence” and are linked inline below; each letter restates the request it answers. Status is as of September 8, 2026; the pending files update this page as they land.
Almost nothing on this site came from a published report, because almost none of it was ever published. It came from freedom-of-information requests — eight to the City of Vancouver and one to the province, plus a data request to the BC Coroners Service — and the story of how the city answered them is evidence in its own right.
Round one: four requests, zero records
In June 2026, four requests went to the city: the waitlist and capacity report the system presumably ran on (A), repeat-registration data (B), staff headcounts and training-program intake (C), and subsidized-access uptake (D).
The four answers arrived as two matched pairs. A and B hit fee walls: $60 for the first, and $2,370 — 161 estimated hours — for the second. C and D received identical same-day refusals citing section 6(2) of the Act, which says a public body isn’t obliged to create a record that doesn’t exist: no compiled staffing counts, no uptake analysis, nothing to give. Whatever one thinks of the legal position, note what it concedes — the operator of the province’s largest learn-to-swim program was asserting, in writing, that it holds no compiled analysis of its own queue, staffing, or subsidy uptake. The B file’s letter went further, in a sentence that became this project’s foundation stone: “there is no report to show repeat registration or re-enrollment.”
Round two: narrow, verify, insist
Fee estimates and refusals are opening positions. The $2,370 file was countered with a request for a raw registration-level extract instead of a built report; the city agreed — “little (or no) fee” — and released 151,568 rows covering four years. The $60 file came back at $0 with a $30 deposit refunded, because the actual work took less time than estimated. Both of the fee-walled files were eventually fulfilled in full for nothing.
The refusals fared differently under checking. The D file claimed no records about Leisure Access uptake in aquatics — but the headline figures sit in the board’s own public March 2025 presentation. A reconsideration went in; the review window has since run out, and the public slide stands as the answer the FOI process wouldn’t give. The C file was refiled in a pared form the office could not call a compilation job (one classification count on one date per year); its response is due at the end of September.
Round three: name the document, ask again
The later requests each name a record the city’s own papers prove exists: the query behind the published “over 8,000 requests” figure; the memo behind the Kitsilano options discussion; the Kerrisdale delay memos; the Resolution B costing work, if any; the 2026 Kitsilano attendance ledger. The first of these (file 2026-641) drew a clarification demand that the request was “overly broad”; it was narrowed and the clock restarted. A named, existing record can’t be refused under s.6(2) — it can only be produced, or withheld under an exemption that itself confirms the record exists. Those files come due through September and October, before the election.
The provincial track
A request to the ministry responsible for sport asked a simple question: does British Columbia fund swim instruction at all? The ministry’s first response was a $150 fee estimate whose framing was itself informative — its suggestion for narrowing pointed to Community Gaming Grants, the generic nonprofit stream, as the place to look. The narrowed request (which converts that hint into a yes-or-no on the record) eliminated the fee entirely and comes due September 18. A parallel data request sits with the BC Coroners Service on child drowning statistics.
What the trail shows
Three things, in ascending order of importance. The FOI office, dealt with in good faith and precise language, behaves reasonably: every fee in this project ended at zero. The records that should have existed mostly didn’t: the system’s management numbers had to be assembled from raw exports by the requester. And the numbers that did exist travelled privately — confidential memos to commissioners, in-camera documents — surfacing only on request. The library holds the City’s response letters (each restates the request it answers) and the released aggregate dataset; the provincial letters will be added when that file closes. None of this needs to be taken on faith.