Inside one plenary: TSG SA#112

Easy to get wrong

One from each chapter that has one. The course carries 19 in all.

• 1.3 Why a group of groups needs a meeting of its own

  Fewer people checked in than registered. That is not evidence of a poorly attended meeting, and it is not evidence of anything else either: the record says how many people did each thing and nothing about why.

• 2.8 SA#112 as arithmetic

  The mistake is not preferring one number. The mistake is quoting either of them without saying which question you asked. Both totals are true sentences about SA#112, and they are not the same sentence.

• 3.6 How to read a plenary agenda

  Sixty-three documents under agenda item 20 is not the same as sixty-three Release 20 documents. Item 20 is the bin for Rel-20 change requests only, while across every kind of document — work items, studies, reports, letters — 140 belong to Rel-20 ^[1]. Confusing the two turns a change-request count into a release size.

• 4.4 The kinds of document, and the ten words that decide them

  26 specifications and reports were approved at this meeting ^[2]. Every one of them still had to be produced by the support team and put on the server afterwards, and nothing in this record says when that happened. "Approved at SA#112" and "available" are different claims and only the first one is evidenced here.

• 5.3 Release planning — the timeline, the checkpoint and the names

  "The plenary approved the Release 21 timeline" is wrong twice over. The word is "agreed", not "approved", and the agreement was reached in the joint session with TSG RAN, not by TSG SA on its own.

• 6.6 The new work — what the plenary let start

  The work item was approved. The legal question was sent upwards, unanswered. Both halves are true at once, and a summary that reports only the approval leaves out the reason anybody remembers the item.

• 7.4 The change requests — 689 of them, in packs

  Neither number is the wrong one. Quoting either without saying which question produced it is what makes a reader unable to check you. Chapter What this record does not tell you lists the other places where this record supports two true sentences at once.

• 8.4 The post — who wrote to 3GPP, and who 3GPP wrote back to

  Noted does not mean ignored and it does not mean agreed. It means the plenary has read the letter and is taking no decision on it at this meeting. Writing "3GPP accepted NGMN's timeline" from a noted letter would be inventing a decision that was explicitly not taken.

• 9.7 Release 20 at this meeting, and what the numbers will not say

  The tempting sentence is "Rel-20 is about AI, management and ambient IoT". What the record actually supports is "at this meeting, more separate change requests were submitted against the AI/ML enablement, management and ambient IoT specifications than against any others". The second sentence survives somebody checking it.

• 10.6 The report is a draft, and it shows

  The temptation is to treat the wobble as trivial and everything else in the report as solid. The correct conclusion is narrower and more useful: this document has not had a proof-reading pass, so a single word in it — a name, a date, a status — is worth checking against the document list before you rest an argument on it.

• 11.4 How to follow a plenary yourself

  A document number is not a decision. Searching a meeting for a title returns every version of it, and the versions have opposite results. Always take the status from the document list, never from the fact that a document exists.

• 12.10 What this record does not tell you

  A generated register is trustworthy about provenance, not about coverage. "Not in the register" means nobody extracted it, not that it did not happen.