Why standards exist, and who writes them · chapter 2 of 4 · 4 minutes
2 Who writes them — two organisations, two habits
3GPP's pyramid of named groups next to the IETF's open participation, and the two very different ways each of them reaches a decision.
2.1 3GPP: a pyramid, and everybody has a name
3GPP is built out of named bodies that report to one another. At the top of the technical side sits a Technical Specification Group, a TSG. A Working Group is an official subgroup of a TSG and reports to it, and sub-groups sit below that again §2.
Work is handed to a Working Group that answers directly to a TSG; in practice it may then be done in a subgroup of that Working Group §1.
The meeting calendar records three of these families with long runs: 3GPP SA with 173 meetings recorded, 3GPP RAN with 160, and 3GPP CT with 110.
Each runs working groups of its own. SA has SA1 to SA6. RAN has RAN1 to RAN5, plus a RAN6 that met 19 times between 2016 and 2020. CT has CT1, CT3, CT4 and CT6 — there is no CT2 or CT5 in the calendar at all. All told, 27 3GPP bodies hold a meeting series of their own [2].
Two more bodies meet above the technical groups: the Project Co-ordination Group, with 64 meetings recorded from 1999 on 3GPP PCG, and the Organizational Partners, with 58 from 2015 on 3GPP OP.
2.2 The people who never leave
Around the delegates sits a permanent staff, called the Support Team or the Mobile Competence Centre — the permanent secretariat of 3GPP §2.
They manage the work of the groups: they edit every specification once it is under change control, they prepare and support the meetings, and they handle correspondence with outside bodies §3.1.
That last job matters more than it sounds. 3GPP is not a legally constituted entity, so when a code point has to be registered with an outside body, the staff register it in the name of one of the Organizational Partners on behalf of them all §3.2.
Each individual document has an owner too. Every specification has one prime responsible TSG, may have secondary ones, and must have a rapporteur: a delegate from a member company who takes part in the responsible group regularly §4.2.
2.3 The IETF: no membership at all
The contrast is sharp. There is no formal membership in the IETF, and participation is open to all [4]. People take part as individual technical contributors, not as formal representatives of the company paying them RFC 2418.
The work is done in working groups, which are collected into areas, and each area is run by one or two Area Directors RFC 2418. Those Area Directors sitting as a body, together with the IETF Chair, are the Internet Engineering Steering Group — the IESG, which approves IETF standards [5].
How a working group is formed, chartered, run and closed is written down in its own document RFC 2418, and the full cast of bodies around the process is named in another RFC 2028.
2.4 The two ways of deciding
The two also reach a decision in completely different ways.
In 3GPP a decision is a two-step act by named bodies. The Working Group agrees a change, and the parent TSG approves it; the verdict on every single document is recorded §4.6.4.
The IETF rejects that shape outright, and has done since the beginning:
Rough consensus is a judgement rather than a tally. If the chair decides an objection has been truly considered and answered, the chair may declare consensus over the objector's continued objection [6].
And it really is not a headcount: 51% does not qualify as rough consensus, 99% is better than rough, and it is the chair who decides [7]. Consensus does not require everybody to agree, though that is preferred [8].
One honest gap: how a 3GPP decision is actually carried, and with what weight per member, is not in the working-methods document at all. Those rules sit one level up, in the 3GPP Working Procedures [9].
Next: what all these groups actually do to a document, one change at a time, in One idea, from proposal to published text.
Where the numbers in this chapter come from
- RFC 7282, section 1 the words of section 1 of RFC 7282 as stored at /var/www/whatthespec.net/data/data/rfcs/rfc7282.txt
- 27 3GPP bodies hold a meeting series of their own distinct body names starting '3GPP' in table `meetings`; the list is every one of them, smallest included, because only the larger ones get an entry of their own in this register, read from /var/www/whatthespec.net/data/database/api/portal_meetings.sqlite on 2026-08-04
- 4,371 of them are 3GPP meetings, from 1998-12-07 to 2028-06-13 rows in table `meetings` whose body name starts '3GPP'; the span is the earliest and latest start date among them, read from /var/www/whatthespec.net/data/database/api/portal_meetings.sqlite on 2026-08-04. The latest dates are meetings already scheduled, not meetings held
- RFC 2418, section 1 the words of section 1 of RFC 2418 as stored at /var/www/whatthespec.net/data/data/rfcs/rfc2418.txt
- RFC 2418, section 1 the words of section 1 of RFC 2418 as stored at /var/www/whatthespec.net/data/data/rfcs/rfc2418.txt
- RFC 7282, section 1 the words of section 1 of RFC 7282 as stored at /var/www/whatthespec.net/data/data/rfcs/rfc7282.txt
- RFC 2418, section 3.3 the words of section 3.3 of RFC 2418 as stored at /var/www/whatthespec.net/data/data/rfcs/rfc2418.txt
- RFC 2418, section 3.3 the words of section 3.3 of RFC 2418 as stored at /var/www/whatthespec.net/data/data/rfcs/rfc2418.txt
- The 3GPP Working Procedures the address as written in the reference list of TR 21.900, reference [8]
Check yourself
Answers appear when you pick one, with where they come from.
Q2.1 In 3GPP, what is a Working Group?
A Working Group is an official subgroup of a TSG and reports to that TSG. Sub-groups sit below it again. §2
Q2.2 Who may take part in the IETF?
There is no formal membership in the IETF and participation is open to all; people take part as individual contributors. RFC 2418, section 1
Q2.3 What does a 3GPP rapporteur do?
Every specification must have a rapporteur, and the role is filled by a delegate from a member company who takes part in the responsible group regularly. §4.2
Q2.4 How does an IETF working group settle a technical question?
51% is not rough consensus and 99% is better than rough; it is the chair's call whether an objection has been properly answered. RFC 2418, section 3.3
This chapter was written against TR 21.900 version 20.0.0, and built from a source register generated 2026-08-04. A newer version of the document may say something else.