GxPlex

Vocabulary history

For one word, entered exactly: the first day this record read a retained text holding it, and, for each source, the days its retained text was read and the days that text held the word. The lookup is built and withheld, and nothing on this page looks a word up.

Withheld

An index of every word of every retained text read answers, for any word, whether a retained text holds it, and that includes the names of people and names this record’s boundary refuses. A key made from the word by a digest does not hide the word from anyone who holds it. No list can set aside the names of all people, so the lookup, its index files and its script are withheld, filed on 2026-10-19 and decided by default pending the chair. The index is built on every build and compared with the store, and its digest is printed below.

The rule

A word is a maximal run of letters, digits and the underscore, letter case kept. The entry is looked up exactly as entered, with the white space around it set aside: no plural, no other ending, no other case and no other spelling is looked up in its place. An entry holding any other character is not one word under this rule and is not looked up.

A run of words is not looked up. Answering for a run of words would need the order of the words in each text, and an index holding that order holds the text itself. This index holds which words each retained text contains and nothing about where they stand in it.

First observed is the earliest day on which this record read a retained text holding the word. It is not the day a publisher first wrote it. Each source’s own first read day is printed beside it, so a first observation on that day marks the start of this record’s reading.

Sources are listed by row identifier and days in calendar order. Nothing is ordered by how often a word occurs, and no word is listed here that was not entered.

What was read

244 retained texts of 86 sources, observed on 98 days from 2026-07-28 to 2026-11-02, holding 294,016 distinct words under the rule above. A retained text is one file named by an observation of one source, counted once for each source whose observations name it; its days are the days of every observation of that source naming it. Of the 132 rows tracked, 46 rows are set aside, each for one reason below, and no absence is filed against them.

The index built

SHA-256 over the index’s lines, one per retained text (its source, its file and its days) and one per word’s key (the retained texts holding it): b2a13d580345b663f6e9c2ae1cb4efaf2e1f710b143752a0fe62b3c4db79b51d. A digest of the whole index answers for no single word.

Last updated: