The mark of the Gazette’s data job now stands on the rows and figures the job filed, on the pages this entry names, and not only on the job’s entries in the change log. On a page the mark links to the rules the job works under, on the method page; on the card a Story draws for the claims floor and on the floor’s share card it is the mark’s words as text. The mark stands on the three things the job files by itself. The weekly pool sample: on Pools, under the leaders’ share, under the three windows and over the table of every label; on Mining, in the one sentence that quotes the share; on the Data Desk’s opening page and on the front page. Strategy’s row in the treasury ledger: on Company treasuries, in the ranking and in the list of records; on Holdings, in the claims floor’s list of holders; and on the front page. The claims floor: on Holdings, under both of its figures; on the Data Desk’s opening page; on the front page; on a Story’s card and on the share card. Some lines that draw on Strategy’s row carry no mark of their own: the totals that sum it with other rows, which are the companies’ total on Company treasuries and the company category’s total on Holdings; a row’s latest movement on Company treasuries, which a person records from the filing; and the line in the section Strategy, filing by filing, that repeats the ledger’s figure when the ledger is ahead of that series. Who filed a figure is read from the change log and from nothing else. An entry written in the job’s template is tied to a figure by what the entry states: the pool sample by the day the entry is dated and the leaders’ share its title states; a company’s row by the ticker and position date the entry is filed for and the amount and date its title states; the claims floor by the percent and block its title states and the sum, in whole bitcoin, its text states. The figure takes what that entry says of who filed it. A figure a person filed carries no mark. A figure the log cannot tie to such an entry, or whose entry is from before 7 October 2026, when these entries began to say who filed them, carries none either: that is not recorded, and nothing is guessed. The downloads of the three datasets gain one field, filed_by, which holds data-job, person, or nothing where the log does not say. In the JSON it stands once on the pool sample, on each company’s row of the treasury ledger, and on the floor and each holder’s row of the claims floor. In the CSV it is a column: the same on every row of the pool sample, and each row’s own in the other two. Each dataset’s metadata file and the data dictionary say what it means and what an empty value means. The field is added from the change log when a file is built for download. The saved files and their methods are unchanged, so each dataset keeps its method version. The three JSON files hold everything the saved files hold and the one field more, the three CSV files gain the one column, and the checksums of all of these files change. No figure changed.
Data Desk