The IETF, and reading the record · chapter 6 of 7 · 5 minutes
6 Two ways of deciding
3GPP decides in two named steps and records a verdict per document; the IETF refuses to vote and hums instead — and where the 3GPP voting weights actually live.
6.1 In 3GPP, two named bodies decide
A 3GPP decision is a two-step act, and both steps have a body's name on them.
The working group agrees a change request. Its secretary collects everything agreed and sends it up. The parent TSG approves it §4.6.4, §4.6.1.
The two words are not interchangeable, and the status list makes the difference explicit: approved means the group in question has the final say, while agreed means the decision still has to be confirmed by a higher body §9.2.
Every document that comes before a meeting ends with one of those statuses recorded against it, and the secretary writes the verdict into the meeting report §4.6.1, §9.2. Across the whole record 97,521 documents ended with the status "approved" [6] and 16,137 with the status "not pursued" [7].
That second word is a deliberate softening. The document says the status "rejected" became unacceptable in some groups and was replaced by "not pursued", and that a reference to one applies equally to the other §4.6.1.
6.2 The word "consensus" does appear in 3GPP
It appears in one telling place. A category B or C change request may not normally be made to a frozen specification — but a TSG may approve one anyway if it is the consensus of the meeting that the exception is justified §4.6.2.
So consensus is available to a 3GPP chair as a way of authorising an exception. It is not the mechanism by which the ordinary decision is made.
6.3 Who may vote, and with what weight, is not in this document
This is the honest limit of the material, and it is worth stating plainly rather than filling in.
The working-methods document describes how a document moves: who proposes, who agrees, who approves, what gets recorded. It says nothing at all about how a disputed decision is actually carried, whether by show of hands, by weighted vote, or by the chair's reading of the room.
Those rules sit one level up, in the 3GPP Working Procedures, which was not among the documents read for this course [8].
6.4 The IETF's founding sentence
The IETF's position is not a procedure. It is a slogan, and it is quoted in an RFC of its own.
Working groups then decide that way as a matter of rule, and the rule is careful to say that consensus does not mean unanimity.
6.5 Rough consensus is a judgement, not a tally
The long form is worth reading once, because almost every misunderstanding of the phrase is answered inside it.
Read what the chair is actually judging. Not how many people are on each side — whether the objection was heard, understood, and answered. A single objector whose point has not been answered can hold the room; twenty people who simply dislike a design cannot.
The other document says the same thing from the other end.
6.6 Humming
Humming exists for exactly this reason. A chair needs to feel the sense of a room without producing anything that can be counted, because a show of hands looks like a vote and gets treated as one RFC 7282.
The hum is a reading, not a result. The chair still has to decide whether the objections have been answered, and that decision is the one that counts.
6.7 Between meetings, both use email
Neither body waits for the next room booking.
3GPP working groups may take decisions by email under rules each group sets for itself, and the document lists an example procedure §8.0. The decision period is two weeks and has to finish a week before the plenary it feeds §8.5. The rule that catches people out is the default.
That is why the same clause requires a delegate who attended the plenary to be able to tell that they have possibly missed a relevant email §8.0, and why a weekly status report is recommended during the period §8.3.
The IETF's Last Call is itself an email procedure, open to anyone, and it is not a fallback for when there is no meeting — it is the normal path RFC 2026.
The two habits look alike and are not. 3GPP's silence-is-agreement runs against a deadline tied to a plenary. The IETF's Last Call invites objections from people who have never been near the working group, and one good objection can stop the document.
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
- 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 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
- TR 21.900, clause 8.6 the words of clause 8.6 of 3GPP TR 21.900 v20.0.0 as parsed under /var/www/whatthespec.net/data/friendlyspec/json/21900/20.0.0
- 97,521 meeting documents ended with the status "approved" rows in table `tdoc` where status='approved', read from /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-04
- 16,137 meeting documents ended with the status "not pursued" rows in table `tdoc` where status='not pursued', read from /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-04
- 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.
Q6.1 What are the two steps of a 3GPP decision on a change request?
The word "agreed" belongs to the working group and means the decision still needs confirming; "approved" is the body with the final say. §4.6.4
Q6.2 Where are the rules on who may vote in 3GPP and with what weight?
The working methods describe how documents move, not how the partnership itself decides. That is a different document. The 3GPP Working Procedures
Q6.3 Is 51% of a working group rough consensus?
Rough consensus is a judgement about whether objections were answered, not a count of who is on which side. RFC 2418, section 3.3
Q6.4 Why does the IETF hum?
A show of hands looks like a vote and gets counted like one, which is exactly what the body is trying to avoid. RFC 7282
Q6.5 In a 3GPP email decision, what does silence mean?
No comment inside the allowed period counts as agreement, which is why delegates are told to watch that they have not missed an email. TR 21.900, clause 8.6
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.