²⁹Reuse and Contribute

We want every statement here to survive being quoted. So cite a claim by its identifier and the status shown beside it, not by the page alone. Pages get reorganised and wording gets revised. The identifier names the exact claim, and the status (confirmed, probable or lead) records what that claim was worth on the day you took it.

Take one confirmed claim: in 1899 the two founders' small sons laid the building's first stone¹. To cite it, repeat the identifier and the confirmed status the site shows against that sentence, and a later reader will see it was offered as settled, not as a guess. This matters most when a claim is not settled. If a finding was probable when you quoted it, your citation should say probable, so that it stays honest if the site later firms the finding up or withdraws it. Whole registers work the same way, one level up. You quote a row in the occupant register, for instance, as that register plus the occupant and year in the row, and a reader can then check it against the scope and source coverage the register prints for itself.

Rights are stated item by item, and we try not to claim more than we know. Each preserved source keeps its archive's rights statement, copied in that archive's own words. Where an archive gave none, the record says so plainly: rights not stated, item held for local research only. That is a caution about what we know, not a legal classification of the item. Read it as the first word on reuse, and not the last.

Two things about the site as a whole are left open on purpose. The licence under which this work may be reused is not yet decided, so the site prints "TBD" instead of letting a reader assume one. The links to the underlying code and data are placeholders until the owner chooses to make the repository public, and where a real address will one day sit there is only a marked stand-in for now. Both are written as undecided deliberately. A licence quietly assumed would be exactly the kind of silent claim the rest of the method exists to prevent.

Last of all, the site asks to be corrected, and it is built to take a correction cleanly. A useful one does three things: it names the statement it disputes, gives a source, and says what that source shows. Every claim on the site already has that shape, and not by coincidence.

A correction sent in that form becomes a proposal. We review it against the existing claim and the evidence behind it. If it is accepted, it replaces the old assessment, and the old wording is kept and marked as superseded instead of being deleted. That is the reason for keeping it: the site can show what it used to say and why it changed. The retired wording also goes into a ledger that stops it from ever slipping quietly back. That ledger of corrected and abandoned claims is the dead-claims deep dive.

A correction need not overturn anything to be worth sending. The same review weighs it whichever way it points. A new source can move a finding from probable to confirmed, or supply a missing date or name, as readily as it can reverse a judgement. A correction never erases the trail behind it, though. Each claim keeps its history, so someone who quoted it at an earlier stage can still see what it used to say and why it moved, and judge whether their own use of it holds. This is why we ask you to cite by identifier and status instead of by page: the page can change under you, but the record of what it once claimed, and on what evidence, does not.

Corrections are welcome because the building will outlast the account of it. Across the four parts of this book it has come through four furniture firms, an occupation, a squat and a threat of demolition, and it still stands on its corner with the first stone legible in the wall¹. This site is one reading of it, made at one time. The record beneath that reading is built to outlive the website that presents it. One task remains: to name the people and archives without whom none of it could have been put together.