3 Study first, rules after — and how to read the trail

3.1 Two kinds of permission

Nothing gets written in 3GPP without a plenary first agreeing that somebody may write it. There are two such permissions and they are not interchangeable.

• A study item, opened by a SID — a study item description — is permission to do homework. What comes out is a study report, a TR, and nobody has to obey it.

• A work item, opened by a WID — a work item description — is permission to write or change binding text. What comes out is a specification, a TS, or a change to one that already exists.

Each is opened by exactly one document, approved at a plenary. That document is the shortest honest statement of what a group promised to do, which makes it the best thing to read when you want to know what a piece of work is actually about.

Here is what that looks like. This is the whole of what SA2 is bound to for the architecture rule-writing, as its work item description opens:

Two things to notice. The work is bound to the conclusions of a named study, by clause number — normative work is rarely allowed to invent. And it is listed key issue by key issue, which is how a study's numbered questions become a specification's paragraphs.

3.2 What the pile is made of

The local record ties 3492 meeting documents to this subject ^[2]. Five kinds account for most of them, and the proportions say more about the state of the work than any single document does.

Kind	How many	What it is
pCR	2036 [3]	proposed text for a study report
discussion	968 [4]	an argument, with no text attached
CR	136 [5]	a change to a published specification
draft TR	41 [6]	a whole report, put up for approval
LS out	38 [7]	a letter to another group

Those five are the kinds the register breaks out. Documents of other kinds are not counted separately.

A pCR is a piece of a study report offered by one company: here is a paragraph, please put it in the report. A study report is built by voting hundreds of those in or out, which is why an area still doing studies is mostly pCRs.

A CR, a change request, is the opposite end. It is the only way a published specification is ever altered: the clauses it touches, the old text, the new text, a reason, and the release it applies to. There are 136 of them here against 2036 pCRs, and What the shape of the paperwork says is about what that ratio means.

3.3 How the trail reads, end to end

The stage 1 work is the only part of this area that has run the whole course, so it is the clearest example of the shape.

• S1. A plenary approves a SID. SP-220084 opened the study at SA#95-e SP-220084.

• S2. The study is revised as it goes. SP-220717 revised it at SA#96 SP-220717.

• S3. The finished report is put to the plenary. SP-230506 approved TR 22.837 at SA#100 SP-230506.

• S4. A plenary approves a WID, opening the rule-writing. SP-230750 did that, at the same meeting SP-230750.

• S5. The new specification is put to the plenary. SP-231398 approved TS 22.137 at SA#102 SP-231398.

Five documents, and between them the whole life of a stage 1 feature. Every other piece of this area is somewhere along the same five steps, and Where the work stands, and why there are no dates says exactly where each one has got to.

3.4 The one word that says what happened

Every document in this record carries a status, and the status is a single word. Approved, revised, noted, endorsed, postponed, withdrawn.

Those words are not interchangeable and the difference between two of them is often the difference between a decision and a delay.

• Approved — the plenary agreed it.

• Revised — it was replaced during the meeting by a different document number. That is not failure; it is the normal way a paper gets fixed while it is in the room.

• Noted — read and filed, no decision taken.

• Endorsed, postponed, withdrawn — agreed in principle, held over, and taken back.

3.5 A study is revised, never replaced

The single most confusing thing about reading a work plan is that a study item keeps its identity while its description changes underneath it. When something about a study has to change — its scope, the documents it targets, the group leading it — a revised SID is approved and the work plan points at the new document.

This area has several such chains, and they are worth reading as evidence that a piece of work was harder than expected.

• The first SA1 study was revised at SA#96 by SP-220717 SP-220717.

• The first radio study was revised at RAN#103 by RP-240799 RP-240799 and again at RAN#105 by RP-242348, which is the revision the work plan still points at RP-242348.

• The architecture study was revised at SA#108 by SP-250833 SP-250833, after SP-250809 was itself revised at that same plenary SP-250809.

• The second radio study was revised at RAN#109 by RP-252819 RP-252819 and at RAN#110 by RP-253246, the one the work plan now points at RP-253246.

3.6 What a version number tells you

3GPP version numbers are three parts, and in this record two of the three are worth reading.

A number below 1.0.0 means a draft: the document exists but has not been approved. TR 33.777, the security study, stands at 0.6.0 TR 33.777, and TS 23.137 at 0.3.0 TS 23.137. Neither is a rule yet.

The first digit, once it is 1.0.0 or more, matches the release. TS 22.137 at 19.1.0 and TR 22.837 at 19.4.0 are Release 19 documents TS 22.137 TR 22.837; TR 23.700-14 at 20.0.0 and TR 23.700-15 at 20.1.0 are Release 20 ones TR 23.700-14 TR 23.700-15.

That single digit is the fastest way to tell whether a document belongs to the finished half of this area or the running half.

3.7 Groups talking to each other

The last kind of document worth knowing is the liaison statement: one group writing to another, usually to ask a question or to warn that something is coming.

38 of the area's documents are outgoing liaisons ^[7]. Three of them the record marks as approved and sent:

• S2-2511045 — SA2 writing about what needed coordinating with RAN S2-2511045.

• R3-261541 — RAN3's reply to it R3-261541.

• S6-254616 — SA6 announcing its own study to whoever else cared S6-254616.

A pair like the first two is the most useful thing in a paper trail: it is the moment where two groups had to agree on a boundary, in writing, with both halves preserved.

3.8 Where to read the real thing

None of these documents is carried here. The four work item descriptions that open the current rule-writing are the shortest way in:

• RP-261566 for the radio RP-261566.
• SP-260297 for the architecture SP-260297.
• SP-260317 for the application layer SP-260317.
• CP-261274 for the protocols CP-261274.

Next, the map: which groups appear in this area at all, and what the document counts say about where the weight sat — The map of the groups, and where the weight sat.