The capital channel has no second publisher

ofcc-projects is the only connector blocked by a policy choice rather than a data problem: the site 404s a self-identifying agent, crates/connect/src/cache.rs refuses to send a browser string, and impersonating one would be discourteous besides.

#11 recorded that and named what had not been tried:

Choose an alternate publisher. Not yet surveyed. The Auditor of State, the Controlling Board, or capital appropriation acts may carry project-level detail.

A blocker whose alternatives are unsurveyed is a blocker nobody has finished reading. This is the survey.

The refusal is the estate’s, not the commission’s. ofcc.ohio.gov answers every path with an identical 5,248-byte 404 — root, robots.txt, sitemap.xml, and a path invented to be absent. obm.ohio.gov does the same at 5,278 bytes; tax.ohio.gov at 5,036, which this repository had already recorded for a different connector.

So it is one filter across *.ohio.gov and not a decision this commission took. Two things follow. No 404 from that host is evidence about any document — the same trap the casino connector fell into and was rescued from by checking a CDN. And asking OFCC to change it would be asking the wrong body. ohioauditor.gov, on its own domain rather than under ohio.gov, answers this project’s agent with a 200.

The capital appropriation acts size the channel and cannot allocate it. They are reachable through the Legislative Information Systems API this repository already speaks for ohio-session-laws, and Am. Sub. H.B. 2 of the 135th General Assembly, as enrolled, carries the Facilities Construction Commission’s school appropriations as exactly two lines:

Public School Building Fund (Fund 7021)
  C23001  Public School Buildings                    $  7,000,000
School Building Program Assistance Fund (Fund 7032)
  C23002  School Building Program Assistance         $600,000,000

Six hundred million dollars in one line item, with no district named. The act does enumerate hundreds of individual local projects by name — a school playground renovation at $250,000, a church school building expansion — but those are community project earmarks in a different section and a different channel, and reading them as school facilities assistance would be a category error that the act’s own fund codes rule out. verified the enrolled text, fetched from search-prod.lis.state.oh.us/api/v2/general_assembly_135/legislation/hb2/05_EN/html/

The Internet Archive holds the site and not the data. ofcc.ohio.gov/data.json has four captures and every one is eighty kilobytes of WebSphere Portal HTML rather than JSON; the annual-reports index carries no document links because the page renders client-side. This is the trap recorded elsewhere in this repository as a lesson: a Wayback 200 on a portal proves the crawler stored a shell, not that it captured anything.

ofcc-projects stays Declared and stays needs-access, with a blocked_on that now says what was tried rather than only what stands in the way.

What is left is genuinely narrower. The acts are ruled out on evidence rather than on a guess, and ruled out for a reason that will not change: a lump appropriation cannot be disaggregated by anyone, including the commission, because the allocation happens after the act in Controlling Board releases. controllingboard.ohio.gov does not resolve at all, which is the one thread this survey leaves live — Controlling Board releases itemize project by project and are the natural home for exactly this, and finding where they are published is the next thing to try.

RESOLVED by #254, and it closes the thread rather than advancing it. controllingboard.ohio.gov never existed. dig returns NXDOMAIN for the apex and for www, curl exits 6 — “couldn’t resolve host” — on both schemes, and the Internet Archive holds zero captures of the domain, ever. The hostname above was a guess made while writing this paragraph, and it read as a lead because it was plausible rather than because anything pointed at it.

The board publishes under obm.ohio.gov — /areas-of-interest/controlling-board/, and its agendas and minutes beneath that — which is the host named three paragraphs above as answering with an identical 5,278-byte 404. It answers both of those real paths with exactly that 404, byte for byte what it returns for a path invented to be absent. The board is not a second publisher; it is the same wall.

The archive holds the portal and not the documents, for the second time on the same platform. 323 distinct URLs are captured under obm.ohio.gov/areas-of-interest/controlling-board, every one of them HTML: IBM WebSphere Portal’s !ut/p/ state-token shell with its Angular and its stylesheets. Not one .pdf, .xlsx or .doc among them. That is the trap recorded above for ofcc.ohio.gov/data.json, reproduced exactly.

So the capital channel’s two remaining routes are not independent, and the paragraph below that calls the Auditor the second live thread is counting one too many. OFCC and the Controlling Board are different bodies with different statutory roles, and they sit behind one filter on one portal platform. “Try the other publisher” is therefore not a strategy here, because the publishers share the infrastructure that is the blocker. What is left is institutional — a records request, or a publisher that serves a self-identifying agent, which so far means ohioauditor.gov and its own domain.

The Auditor is the second live thread and a better-founded one than when #11 named it: it is already a wired publisher here, and it answers this project’s agent. What its district audits carry is the district’s own capital spending rather than the state’s contribution to it, so it is a different quantity and not a substitute — but it is reachable, which none of the others is.

Send a browser string. Rejected again, and now for a second reason. The first is the standing one in cache.rs. The second is that the filter is estate-wide, so the decision is not about one agency’s preference and changing it for OFCC would change it for every Ohio source at once.

Treat the $600m line as the capital channel. Rejected. It is the channel’s size, and this repository already declines to publish a statewide figure it cannot put beside a district — the same reasoning that keeps the scholarship channel’s aggregate from standing in for its incidence. A per-district capital figure is the thing that is missing and a state total is not a version of it.

Ask OFCC to serve a self-identifying agent. Rejected as addressed to the wrong body, per above.