The source

Minutes 7–13. One printed entry, and what you would extract from it.

Terms on this page

• OCR: optical character recognition: software that turns a scanned page into text; its mistakes are the “garbles”.
• entry: one printed biography in the Record of Services; the same person has one in each edition.
• schema: the agreed set of fields a record has.

All terms: glossary.

The Record of Services

From 1883 every Colonial Office List carries a section headed, in most years, “Record of the Public Services of Officers”: between 1,400 and 2,900 entries per edition, cumulative, each a compressed career. Earlier editions, from the first one the project used (1867), carry a shorter “account of the services of the principal officers of the several colonial governments”; Wodehouse’s 1867 and 1878 entries come from that section. The List itself began in 1862. Here is the 1889 entry for the director of the botanic gardens at Peradeniya:

TRIMEN, HENRY, M.B. (Lond.), F.R.S., F.L.S. —Formerly lecturer on botany at St. Mary’s Hospital, London, 1867 to 1875; was senior assistant to department of botany, British Museum, 1869 to 1879; director Royal Botanical Gardens, Ceylon, Feb., 1880; author of numerous works and papers on pure and applied botany.

The same man has an entry in seven editions, 1883 to 1896, and they differ (all seven). Later entries are denser:

BALMER, F. E.—Dep. chief acctnt., G.P.O., E. Africa Prot., Mar., 1919; ch. acctnt., G.P.O., Kenya, Oct., 1919; ch. acctnt., Kenya-Uganda-Tanganyika post and tel. service, 1933; senr. dep. P.M.G., do., 1935.

CONNOLLY, HON. SIR JAMES DANIEL, KT. BACH. (1920). —B. 1870; ed. Warwick Pub. Schl. and Catholic Schl.; M.L.C., Western Australia, 1902; col. sec. and min. of com. and lab., 1906; col. sec. and min. of pub. health, 1910; leader of leg. coun. from 1906 to 1911; M.L.A. for Perth, 1914; agt. genl. in London for Western Australia, 1917-23; agt. gen. in London for Malta, 1930.

Mark it up yourself

Before the model does it, do it by hand. Open the mark-up exercise (or the paper handout in the room) and choose Balmer (entry 2). Tag the last two positions, their places and their years. Three minutes, then one minute to compare: what does “do.” refer to, and where should that decision be recorded? Press Compare to see the saved pipeline record. The other entries are for further practice.

Three things come up every time, and they are the structuring decisions of the next step:

• “do.” and a clause with no place: does the posting inherit the previous one’s place?
• “col. sec. and min. of com. and lab., 1906”: one event or two?
• “KT. BACH. (1920)”: one honour with a year, two honours, or a title?

The scale

Volumes OCR’d (Colonial Office List family, 1867–1966)	85
Editions with a Record of Services	every one from 1883, except the slim 1946 restart
Entries per edition	1,400–2,900
Resolved people (Colonial Office List)	27,526
Career events	180,470

The India Office List is the same shape and was processed by the same scripts, rerooted to a second corpus: 19,400 people, 124,694 events.

The published description mentions Britannica OCR tools. This session begins with existing OCR and does not demonstrate those tools. The volumes used here were OCR’d with olmOCR on a Compute Canada cluster. The same structuring and review steps can be tried on text from another OCR workflow. What matters here is that OCR output is the input, with its garbles (“ABERY” for CARBERY, “ACARLTHUR” for McARTHUR), and that a record retains its volume, page and block reference. The workshop packet contains OCR excerpts; a source-note link is not a scan. Check the page image before treating a surprising OCR reading as what the book printed.

What the source can tell us

These are administrative accounts of careers. They make offices, promotions and honours visible through the categories of colonial government. They are not a census of everyone who worked for, lived under or resisted those administrations. Repeated entries can repeat an error as well as confirm a fact.

The map adds another interpretation: a posting becomes a point, often at a capital, and successive postings become an arc. That point is a plotting location, not evidence that the official lived there; the arc is not an observed travel route. Keep the printed wording, the inferred location and the reason for the inference separately.

Why this example

• The entries are formulaic enough that a schema is obvious, and irregular enough that every tenth one breaks it.
• The same person recurs across editions, so identity has to be resolved before anything can be counted.
• The places are nineteenth-century polities, not modern cities, so grounding is a historical decision, not a lookup.
• There is a payoff at the end that depends on all three being right.

Next: Structure.