CT1 — the group that writes what your phone says

Easy to get wrong

• 1.3 What CT1 is, and the three names it answers to

  Searching for "CT1" in the wrong store finds nothing. This is not a naming quirk to smile at — it is the single most common reason somebody concludes a document or a meeting does not exist when it does. Match the store: C1 in the catalogue, C1- in front of a document number, "3GPP CT 1" in the calendar.

• 2.4 Stage 3 — the difference between a plan and a protocol

  "Stage 3 is the detailed version of stage 2" is the mistake that wastes the most time. It leads people to read the architecture document, form a picture, and then argue with the protocol document when the picture does not fit. The protocol document is not elaborating on the architecture — it is answering a different question, and it is the one an implementation has to obey.

• 3.2 What CT1 owns — 229 documents and the six that matter

  "Active" is a flag in the catalogue, not a promise that anybody is working on the document. A specification can be active, have had no change proposed to it for years, and still be the thing an implementation has to obey. Do not read the 229 as 229 live pieces of work.

• 4.3 TS 24.501 — two protocols living in one document

  The NAS has no link of its own. It is glued together out of two other connections that other documents define, which is why a NAS procedure can fail for reasons that have nothing to do with NAS — the radio side dropped, or the leg between the base station and the core did. The document handles this explicitly, as "lower layer failure", in the abnormal-case subclause of nearly every procedure.

• 5.3 Finding your way around TS 24.501

  Reading clause 5 without clause 4 produces confident wrong answers. A registration procedure in clause 5 will say the device applies unified access control before sending — and the rules for what that means, which category applies and what happens when access is barred are entirely in clause 4.

• 6.4 Getting on the network, and getting a connection through it

  A 5G-GUTI is not an account number. It is deliberately replaced, and code that treats it as a stable key for a subscriber will break the first time the network reassigns one — which the specification requires it to do on every successful initial registration.

• 7.4 What a NAS message looks like on the wire

  An information element identifier is not a global name. The same hexadecimal value means one thing in a registration request and something else in a service accept, so a parser that keeps one table of identifiers for the whole protocol will decode fields into the wrong shape. The identifier is only meaningful inside its message.

• 8.3 From a document number to a published version

  A document number and a change request number are different things and neither can be worked out from the other. C1-262074 is a document filed to a CT1 meeting; 7174 is the change request number against TS 24.501; CP-261125 is the document under which the change reached the meeting recorded in the change history. Three numbers, one change, and the change history only holds two of them.

• 9.3 One CT1 meeting, by the numbers

  The C1-161 report is a draft ^[3]. Its numbers can change before it is approved. The report of the meeting before it, C1-160, is approved ^[4]. Anything said about C1-161 has to carry the word "draft" or not be said — and that includes the participant count and the work-item count in this chapter.

• 10.2 Who CT1 writes to, and what the letters count

  Both tables stop at four on purpose: the register carries the four busiest partners in each direction and no more. A group missing from them has no entry at all, so its absence is not a zero and no number may be quoted for it. This is the single easiest mistake to make with this data.

• 11.2 The work items that fill a CT1 year

  A work item name is not an abbreviation the register can explain. "TEI19", "5GSAT_Ph3_ARCH" and "MINT_Ph2" are labels a document carried; anybody translating them is using knowledge from somewhere else. Where a specification defines the term in its own text, that is a citable source — and where it does not, the honest answer is that the name is just a name.

• 12.1 The neighbours — and how little the record says about them

  If you came here to find out what CT4 does, this data cannot tell you and this course will not guess. The place to look is the specification status report, which lists every specification with the group that owns it ^[5], and the 3GPP work plan ^[6]. Both are addresses the register carries.

• 13.3 Following CT1 yourself

  The six counted words do not add up to the 617 documents at that meeting ^[7], and the two uncounted words are why. Never work out a missing outcome by subtracting the ones you have — the answer will be wrong and will look convincing.

• 14.3 What this record cannot tell you

  Both mistakes produce numbers that look derived rather than invented, which is exactly what makes them dangerous. A number nobody counted is not more trustworthy for having been calculated from two that somebody did.