10 The report is a draft, and it shows

10.1 What the report is

The secretary writes it. At SA#112 that was Maurice Pope of MCC ^[2], and the result is the file SP-112_Draft_Report_v005 ^[1].

It is the only document of the meeting that tells you why anything happened. The document list gives you a status word; the report gives you the sentence the chair said out loud. Both of the meeting's long arguments — the training data sent up to the Project Co-ordination Group ^[3] and the guidance that closed the modernisation study ^[4] — exist only here.

Which is exactly why its state matters.

10.2 Its shape

Two halves.

The narrative walks the agenda in order, item by item and sub-item by sub-item, from "The agenda was reviewed and approved." ^[5] at the start to "The meeting was then closed." ^[6] at the end. Where an item drew nothing, it says so: the sentence "There were no contributions to this agenda item." appears 40 times ^[7].

The annexes are tables. One lists every document; the others list the attendees, the letters, the approved specifications, the approved work items and the change-request packs. Each table closes with a count of its own rows, which is where a large share of this course's figures come from.

10.3 What the annexes count

Annex table	Rows
documents	346	[8]
registrations	568	[9]
checked in	365	[10]
registered online	110	[11]
incoming liaison statements	18	[12]
outgoing liaison statements	2	[13]
specifications and reports approved	26	[14]
specifications and reports for information	6	[15]
change-request packs	120	[16]
new work and study items approved	30	[17]
work and study items updated	11	[18]
work item exception sheets	0, "No entries found."	[19]

Two of those can be checked against the meeting's document database and both agree exactly: the document total and the change-request pack total. That agreement is the strongest evidence in this course that the paperwork of SA#112 hangs together.

10.4 It is a draft and it says so

The file is version 0.0.5 and its own cover page reads "Status: Draft - for comment".

That is not a technicality. It means no sentence built on this document may call anything final, confirmed, ratified or the official record. The wording that survives scrutiny is "the draft report of the meeting records that…", and this course uses it on purpose, once a chapter, however clumsy it reads.

There is a second gap. The chair's final notes are published after the meeting closes — the agenda itself says so ^[20] — and they are not on this machine. Nothing here can say what they contain, and it is quite possible that they settle something this draft leaves open.

10.5 The place it contradicts itself

Under agenda item 1.5 the narrative reads: "The report of meeting SA#111 was approved." ^[21]

The status line under that same document, in the same report, reads "Not concluded."

And the meeting's document database records the document itself, SP-260362, as approved.

Three statements about one thing, two of them in the same file, and this course is not going to pick the one that reads best. Report all of them or none. If you need to know what really happened to the SA#111 report, this record cannot tell you and the honest answer is to wait for a version of the report that is not a draft.

10.6 The place it has not been proof-read

The chair's surname appears in the report as both "Jaine" and "Jain".

That is a small thing and it is a useful one, because it sets the right expectation. A draft report is a working document typed at speed during four days of meetings. Spelling in it is not authoritative, and neither is anything else that looks like a detail somebody would fix on a second pass.

10.7 What it never records

Three things are absent by design rather than by accident, and no amount of careful reading will recover them.

• Vote counts. There is no vote count anywhere in this report. 3GPP works by consensus, so what gets written down is the decision and, where somebody objected, the objection — "Qualcomm objected to this decision." ^[22]

• Attendance per company. The report counts people — 568 registered, 365 checked in — not positions.

• Reasons. It records what was decided, not why each participant wanted it. Where a reason appears at all it is because somebody said it out loud and the secretary wrote it down, as with the terms of reference ^[23].

From that follows a rule this course keeps: a decision at SA#112 may never be described as unanimous, widely supported or controversial. The record does not carry any of those.

10.8 What it leaves out entirely

Only 284 of the 346 documents get a paragraph of their own ^[24]. The other 62 appear only as a row in an annex table.

For those 62, the row is the whole evidence: a number, a title, a source, a status word. If you want to know what was said about one of them, the answer is that nothing was — or at least that nothing was written down.

10.9 How to quote it

The wording is not fussiness. Each of these pairs is a real difference, and the right-hand version is the one that survives somebody opening the file.

Do not write	Write instead
the official record of SA#112 states	the draft report of the meeting records
SA#112 confirmed	the draft report records that … was approved
the meeting agreed the Rel-21 timeline	the timeline was agreed in the joint session
the report of SA#111 was approved	the narrative says approved and the status line says "Not concluded."
the chair, Puneet Jaine	the chair, Puneet Jain — the draft spells it both ways
the decision was contested	Qualcomm objected, and the record says nothing about anyone else

Two habits go with the table. Name the document version you read — this course read version 0.0.5 — and prefer the document list for a status and the report for a reason, because that is what each of the two is good at.

10.10 What is not the report

There is a summary of this meeting on whatthespec.net written by a language model ^[27], and 224 of the meeting's documents have a machine-written summary of their own ^[28].

They are genuinely useful for finding your way into 346 documents, and this course would have taken far longer without something like them. They are also not evidence. A machine-written summary may be cited as a summary and may never be the source of a claim about what a meeting decided.

10.11 Where this leads

You now know what the record is worth. Part 4 is about doing this yourself: the five files every plenary leaves behind and the order to open them, in How to follow a plenary yourself, and then the honest list of what none of it will tell you, in What this record does not tell you.