Summary

  • APNIC's 1998 annual report is the only document in the reviewed record that attaches a day to Bruce Campbell's arrival: "commenced 16 December 1998", printed beside the appointment itself.
  • RIPE NCC's annual reports for 1999 and 2001 print the size of the workforce as a number and, in the text reviewed, no named staff line; the five surfaces on the RIPE NCC side that do name him print an organisation, in two cases a title, and never a start date.
  • The only RIPE NCC start period in the record comes from a third-party mirror of a self-reported profile, and that same page places his APNIC start eleven months earlier than APNIC's own printed day.

The line with a day in it

APNIC's Annual Report 1998 carries a short appointments entry, and it is the only place in this whole record where a career acquires a calendar date. The provider-reported excerpt of the PDF reads: "Gerard Ross, Technical Writer (commenced 14 December 1998) / Bruce Campbell, Systems Administrator (commenced 16 December 1998)" (APNIC Annual Report 1998).

Read the shape of it before the substance. It is not a staff register. It is a sentence inside an annual report, the kind of line a secretariat writes when it wants the year's arrivals visible: the name, the title, the day the person started. A second, separately reported staff reference in the same report calls the same man "System Administrator" — one word shorter, in a list of current staff, with no date at all. The day appears exactly once, in the appointments line, and the report's own two spellings of the title travel with it.

That is a small thing to build a briefing on, unless you go looking for the second date and discover there isn't one.

The genre that counts

RIPE NCC published annual reports in the same years, and in the retrieved text they do not carry appointments lines. They carry arithmetic. The 1999 report states: "The RIPE NCC had 59 employees as at December 31, 1999." (RIPE NCC Annual Report 1999, ripe-207). The 2001 report states: "In 2001, the RIPE NCC had a workforce of 75.9 FTEs (Full Time Equivalent) versus 61.7 in 2000." (RIPE NCC Annual Report 2001).

The 2000 report is a different problem. Its landing page says only this: "This RIPE Document is only available in PDF format." (ripe-218). That PDF was not opened, so the 2000 report is unread here, not empty, and I will not pretend to know which genre it used.

What the two readable reports do show is a genre shift inside one publisher. The counted noun changes from employees at a stated date to full-time equivalents across a year, and the unit of account changes with it. Neither sentence is obliged to name anyone, and neither does. A headcount is not a defective staff list; it is a different instrument, built to describe scale rather than to identify people, and it does that job with figures no name could improve.

The cost is specific. If you want to know when a particular engineer joined the RIPE NCC, the annual-report genre will not tell you, no matter how carefully you read it.

Five surfaces, no start date

So the record comes from somewhere else, and it says less than the reports did. Three meeting attendance registers name him: the RIPE 41 list, last updated 18 January 2002, prints "Bruce Campbell | EU | RIPE NCC" (RIPE 41 attendee list); the RIPE 43 list, last updated 13 September 2002, prints the identical row (RIPE 43 attendee list); the RIPE 48 list, last updated 31 May 2004, reorders its columns and prints "Campbell | Bruce | - | RIPE NCC" (RIPE 48 attendee list).

Two more surfaces are written by his own hand. A 12 July 2002 message to the RIPE DNS working-group list signs off "Bruce Campbell / Systems/Network Engineer / RIPE NCC" (RIPE DNS WG archive). A 7 August 2002 post to the rt-users list signs off "Bruce Campbell / Systems/Network Engineer / RIPE NCC / Operations/Security" (rt-users archive).

Five surfaces, three from the publisher and two from the man, and the same absence in every one of them: no start date, no interval of service, no reporting line. The attendance rows are registration data — a row in a delegate list says who registered for a meeting, not who employs whom — and it would be a mistake to promote them into a payroll record. My own arithmetic, printed by nobody: from the last update of the earliest of those registers to the last update of the latest is 2 years, 4 months and 13 days, and across that whole span the record never once dates him.

The only date at the second employer is his own

One surface does give the RIPE NCC period a start. It is a third-party mirror of a self-reported professional profile, and what it prints is month-level: "Jan 2001 - Jan 2005 (4 years) in Amsterdam" (profile mirror). Not an employer record, not an administrative register, and the original profile page was not inspected here. It is the only RIPE NCC start anywhere in the set.

That same page dates his APNIC period from January 1998. APNIC's own report says he commenced on 16 December 1998 (APNIC Annual Report 1998). Mine, not any source's: if the profile's month is read as its first day, the two accounts of the same appointment sit 11 months and 15 days apart.

The APNIC side has one more item, and it is a self-authored one: a March 2000 post to the apnic-talk list carrying the signature line "Systems Administrator / Regional Internet Registry / Asia Pacific Network Information Centre" (apnic-talk archive, March 2000). Role and organisation, no date. So on that side there are two surfaces — one employer document that dates him, one signature that does not — and on the RIPE NCC side there are five, none of which dates anything.

Who dates a career

Three candidates offer themselves as the authority for the dates of a working life: the institution that prints the day, the institution that prints the aggregate, and the person who types it himself.

The institution that prints the day is the strongest witness, because it is describing an act it performed, at a precision it chose, in a document it published under its own name. The institution that prints the aggregate holds authority over scale and none over individuals; asking its figures for a name is a category error, not a gap. The person typing his own history controls the telling but carries no corroboration, and the eleven-month divergence above is what that costs: when a self-description is the only dated record, its error becomes the record's error.

That is the real finding, and it is about paperwork rather than people. Two registries, one career, two incompatible genres of personnel evidence — and the one document that dates the man is a sentence in an annual report that nobody would think to cite for a biography.

What this record does not say

Every PDF quotation above is a provider-reported excerpt of retrieved text, not a hand-opened page. The 2001 report's retrieved text may be truncated, so the absence of a named staff line there is an absence in the text reviewed, not a proof about the whole document. The 2000 report is unread. The profile mirror is a third-party copy of a self-description at month precision. Attendance registers record registration, not employment. And nothing in the reviewed set states why any transition happened, or offers any basis for a judgement about competence, conduct or motive; this article offers none.