2 The shape of 3GPP, and where the data stops

2.1 Three layers, and what they are called

The top technical layer is a Technical Specification Group, always written TSG. Under a TSG sit Working Groups, and under those, sub-groups §2.

The definitions clause is short and worth having exactly.

The reporting line is the whole point: a task is handed to a Working Group that answers directly to a TSG, and in practice the work may then be done in a subgroup of that Working Group §1. Nothing in this course's document describes what happens inside those subgroups — it says so itself, and leaves their internal working to them §1.

One piece of vocabulary saves a lot of squinting later. The document uses the bare word "group" as a technical term: a group is a TSG or a TSG sub-group, and a TSG sub-group is a working group, or a subgroup of a working group, or a subgroup of one of those §2.

So when a clause says "the responsible Group shall", it is deliberately not saying at which layer. Whoever holds the document decides.

2.2 The groups that hold meetings

The calendar on this machine records meetings for many bodies, not 3GPP's alone. Of everything in it, 4,371 entries are 3GPP meetings, the earliest the calendar holds starting 1998-12-07 and the latest planned for 2028-06-13 ^[1]. That first date is where the calendar begins; it is not evidence that no 3GPP body met before it.

Four series sit at TSG level. SA has 173 meetings recorded 3GPP SA, RAN has 160 3GPP RAN and CT has 110 3GPP CT. A fourth, GERAN, has only seven recorded, the last of them in 2016.

Below them:

• SA has six working groups with their own meeting series, SA1 through SA6 3GPP SA 1, 3GPP SA 2, 3GPP SA 3, 3GPP SA 4, 3GPP SA 5, 3GPP SA 6.

• RAN has RAN1 through RAN5 in their own right 3GPP RAN 1, 3GPP RAN 2, 3GPP RAN 3, 3GPP RAN 4, 3GPP RAN 5, plus a RAN6 that met nineteen times between 2016 and 2020.

• CT has four: CT1, CT3, CT4 and CT6 3GPP CT 1, 3GPP CT 3, 3GPP CT 4, 3GPP CT 6. There is no CT2 and no CT5 in the calendar.

That whole list comes from one reading of the calendar, which shows 27 3GPP bodies holding a meeting series of their own — GERAN's seven meetings and RAN6's nineteen included ^[2].

Two bodies sit above the technical work and also meet: the Project Co-ordination Group, with 64 meetings recorded from 1999 on 3GPP PCG, and the Organizational Partners, with 58 recorded from 2015 on 3GPP OP.

A further 94 meetings are recorded under the bare name "3GPP" — workshops and gatherings belonging to no single group 3GPP.

2.3 The permanent staff

Delegates come and go with their companies. The Support Team does not: it is the permanent secretariat, also called the Mobile Competence Centre §2.

Its job is wider than minute-taking. It manages the work of the TSGs, edits every specification once the document is under change control, prepares and supports the meetings, and handles contact with outside bodies §3.1.

That editing role matters later. Once a document is under change control the delegates propose changes and the Support Team makes them, which is why a change to a 3GPP specification always goes through the same pair of hands.

2.4 Who owns a document

Every specification has one prime responsible TSG. It may have a prime responsible working group, and it may have secondary responsible groups as well. It must have a rapporteur: a delegate from a member company who takes part in the responsible group regularly §4.2.

The document also carries a model for how work flows between groups, and it uses TSG SA as its illustration: SA1 writes the requirements, SA2 turns them into an architecture and cuts that into pieces, and the pieces go out to the groups that will specify them in detail §6.6.

The document is careful to call this a reference model rather than a rule, and says it can be extended to any TSG or combination of them §6.6.

Two more things sit in that ownership list and are easy to miss. A specification may also have a prime responsible working group, and it may have a prime responsible sub-group below that working group, if the working group decides so §4.2.

And a document is either a Technical Specification or a Technical Report §4.2. The working-methods document uses "specification" as the umbrella word for both, which is why a sentence about "specifications" in these rules is often about reports as well §2.

2.5 Where the work sits

Counting specifications by the group code they carry gives a rough map of where the weight of the work sits.

Group code	Specifications carrying it
S5	685 [3]
R4	458 [4]
S3	343 [5]
S2	321 [6]
S4	303 [7]
S1	256 [8]
C1	237 [9]
C4	236 [10]
RP	200 [11]

Of the whole catalogue, 3,599 specifications are marked still maintained ^[12] and 329 are marked no longer maintained ^[13]. The paper all those meetings generate is on another scale again — 1,629,581 meeting documents are on record ^[14].

2.6 What this data cannot tell you

Four gaps, and they are worth knowing before somebody quotes a number at you.

The working-group list above comes from the meeting calendar, not from a specification. The working-methods document names SA1, SA2 and SA3 in passing while explaining the work model, and mentions CT and RAN; every other group name here was read out of the calendar §6.6.

The calendar runs to 2028 because it holds scheduled meetings as well as held ones, so its last date is never "the most recent meeting". It does not reach back evenly either: SA and RAN start in 1998, CT's first recorded meeting is 2005 3GPP CT, 3GPP SA. Those groups are older than their entries — a first recorded date is not a founding date.

And nobody is named. The Organizational Partners meet, and the document talks about them, but no file here lists who they are.