School of Specs IMS — the call as an internet sessionIn depth

What is written down · chapter 12 of 14 · 11 minutes

12 Maintenance is the main event: change requests and protocol alignment

What a mature 3GPP subject actually consists of — 3509 change requests, one work item reopened in every release since Rel-7, and the outcome words that mean less than they look like.

12.1 Why a subject this old is mostly repair work

A subject in 3GPP has two lives. In the first, somebody writes a new specification and everybody argues about what should be in it. In the second, which is much longer, the specification exists and people change it — a sentence at a time, release after release.

IMS has been in its second life for most of its existence, and the paperwork shows it. Of the 4939 IMS meeting documents on record [3], 3509 are change requests [4].

A change request is the only way a published specification is ever altered: one document, one proposed edit, brought to one meeting.

That single figure is the honest shape of the subject. Anybody who imagines 3GPP's IMS work as architects designing a system is imagining the first life. The second life is a very long queue of small corrections, and this chapter is about what that queue looks like from the outside.

12.2 What the 4939 documents actually are

The store records a type for each document, and the types fall into three groups: edits to a published rule, text proposed for a study, and the paperwork of running the work at all.

Type How many What it is
CR 3509 [4] one proposed edit to a published specification
pCR 767 [5] proposed text for a study report, which is not published yet
CR pack 175 [6] several change requests brought as one bundle
LS out 45 [7] a letter this work sent to another group
WID revised 40 [8] a rewritten work item description
draftCR 38 [9] a change request against text that is still being drafted
WI exception request 19 [10] a request for an exception to what the plan said

The last row is the most human line in the whole record. Nineteen times, somebody had to file one.

Below those seven there is a long tail of smaller kinds. 90 documents are filed as "other" [11] and 61 are discussion papers [12].

51 are drafts of a study report [13] and 32 are drafts of a specification [14].

32 are letters received from another group [15], 30 are status reports [16] and 16 are copies of the work plan itself [17].

None of those changes a rule. They are the noise a working group makes while deciding what to change, and they are worth knowing about only so that a search result full of them is not mistaken for activity.

12.3 The outcome words, and how much they are worth

Each document leaves its meeting with a word written against it. The store carries the word and no definition of it, so what follows is the ordinary English sense and nothing more. Five words carry most of the traffic.

  • revised — 1942 IMS documents [18], more than carry any other word in this record.

  • agreed — 1736 [19].

  • approved — 379 [20].

  • postponed — 297 [21].

  • withdrawn — 189 [22]: taken back by whoever brought it.

Now the part that matters, and it is the single most common mistake made with this data.

The word is about that document and nothing else. It says what happened to one piece of paper at one meeting. It does not say the work item is finished, does not say the specification changed, and does not say the room agreed with the idea — a change request can be agreed for one release and the same idea rejected for another the same afternoon.

The record also holds no reason for any of these words. No store on this machine says why a room settled on postponed rather than agreed. If you need that, you need the meeting report, and this course does not carry one.

12.3.1 The words below the five

Eight more outcome words appear, and each one describes a different way a piece of paper can end its meeting.

  • noted — 173 documents [23]: taken note of.

  • merged — 68 [24]: folded into another document.

  • not treated — 51 [25]: not dealt with at that meeting.

  • rejected — 35 [26], against the 189 withdrawn above [22]. Being taken back by its own author is commoner here than being turned down, which is what a body that works by consensus looks like from the outside.

  • endorsed — 24 [27], not pursued — 20 [28], available — 19 [29], and partially approved — 6 [30].

That last word is the one to remember, because it is the only one that admits a document can be half-accepted. Six IMS documents carry it.

12.4 One job, fourteen entries, every release since Release 7

The clearest picture of what maintenance means here is a single work item that has never been allowed to close.

It is called IMS Stage 3 IETF Protocol Alignment IMSProtoc. Its job is to look at what the IETF has done to SIP since last time and write down whether a 3GPP network supports it. That job never finishes, because the IETF never stops, so 3GPP reopens the work item every release.

Acronym Release Documents
IMSProtoc Rel-7 9
IMSProtoc2 Rel-8 248 [31]
IMSProtoc3 Rel-9 32 [32]
IMSProtoc4 Rel-10 19 [33]
IMSProtoc5 Rel-11 248 [34]
IMSProtoc6 Rel-12 209 [35]
IMSProtoc7 Rel-13 364 [36]
IMSProtoc8 Rel-14 and Rel-15 169 [37]
IMSProtoc9 Rel-15 123 [38]
IMSProtoc16 Rel-16 72 [39]
IMSProtoc17 Rel-17 26 [40]
IMSProtoc18 Rel-18 45 [41]
IMSProtoc19 Rel-19 62 [42]
IMSProtoc20-CT Rel-20 19 [43]

Two habits of naming are visible in that column, and both will trip up a search. The numbering ran as a phase count up to IMSProtoc9 and then jumped: from IMSProtoc16 on, the number in the name is the release the row is marked with.

And the newest one carries a group on the end — IMSProtoc20-CT — because the work is CT's.

The first entry opened with CP-060101 at the CT plenary's thirty-first meeting CP-060101. Nine documents carry it. Every entry after it is larger, because by then the specification existed and there was something to change.

12.5 The same two sentences, Release 8 and Release 20

The best evidence that this is one continuous job is what the proposers wrote. Here is the opening of the justification of the Release 8 phase, from CP-080794 CP-080794:

And here is the opening of the justification of the Release 20 phase, from CP-260044 CP-260044:

Two sentences, one word apart: "provides" became "provided". That is the whole of the change in the stated reason for doing this work between those two releases.

It would be easy to read that as laziness. It is better read as a fact about the subject: the reason has not changed, because the situation has not changed. Release 5 defined IMS, the IETF keeps developing SIP, and somebody has to keep the two in step.

12.6 Where the edits land

The queue has a destination, and it is overwhelmingly one document. 1680 IMS documents name TS 24.229 as the document they change [44] — the protocol specification, the one that says which parts of SIP a 3GPP network uses.

Every one of the protocol alignment entries above declares it as a document it will change, and most of them declare TS 23.218, TR 24.930, TS 29.163 and TS 29.165 alongside it IMSProtoc9.

The pattern is always the same: change what the protocol says, then change everything downstream that quoted it.

The work item descriptions say so in as many words, and the wording barely moves from one release to the next. This is what the Release 15 round told each of its six documents to do IMSProtoc9.

Document What the work item says it will do to it
TS 23.218 review the additional capabilities the IETF has provided in SIP, and document whether they are supported — CT1's responsibility
TS 24.229 the same review, also CT1's
TR 24.930 correct any errors in this report
TS 29.163 changes consequent on the changes to 24.229 made under this work item
TS 29.165 the same, and CT3's responsibility
TR 29.949 the same again

The last three rows exist only because TS 24.229 changed. That is the shape of maintenance in a document family: an edit to the protocol is never one edit.

The rest of the document map — what those numbers are and what else is in it — is The specification family: 241 documents, and the two that are IMS.

12.7 Two numbers that look like a status and are not

The completion percentage. 339 of the 345 IMS work plan rows stand at 100% [45]. That field is a percentage somebody types into the work plan by hand. It is not a status field, nothing recalculates it, and the register holds no status field for a work item anywhere.

The undated documents. 237 IMS documents carry no upload time at all [46]. That is why counting a year and counting the whole record answer different questions, and why the per-year figures never add up to the total. It is also a reason never to build an argument on a trend.

Both of those are read correctly in What is running now: Releases 18 to 20, where the present tense of this subject is set out, and the traps around them are collected in Reading the record yourself, and what it cannot tell you.

12.8 What a change-request count does and does not measure

The counts in this chapter and in The specification family: 241 documents, and the two that are IMS are counts of edits. They are not a measure of importance, and they are not a measure of difficulty. Edits happen where text is dense and where people disagree about detail, which is not the same as where the interesting ideas are.

Three readings are safe, and they are the only three.

  • C1. A high count on a long-lived protocol document means the document is being kept in step with something outside 3GPP. TS 24.229's 1680 is that [44].

  • C2. A high count on a document belonging to another subsystem means this work genuinely reaches into it. TS 29.562's 230 IMS documents are that TS 29.562.

  • C3. A count of zero on a document that work items declare means nothing arrived in the window this store covers. TS 23.218 and TS 22.173 are that TS 23.218 TS 22.173.

What no count supports is a sentence comparing this subject with any other area of 3GPP work. Nothing in this register counts positioning, sensing or ambient IoT as subjects of their own, so IMS is not "the most maintained" anything on this evidence.

Where the numbers in this chapter come from

  1. why IMSProtoc2 was proposed the opening sentences of the justification of the work item description /var/www/whatthespec.net/data/data/wis/360029/CP-080794/CP-080794.md, copied word for word, read 2026-08-05. That file was taken as this work item's own because its name begins with CP-080794, the document the work plan points at
  2. why IMSProtoc20-CT was proposed the opening sentences of the justification of the work item description /var/www/whatthespec.net/data/data/wis/1110010/CP-260044/CP-260044.md, copied word for word, read 2026-08-05. That file was taken as this work item's own because it states the unique identifier of this work plan row
  3. 4939 IMS meeting documents rows of table `tdoc` whose work item column holds, as a whole word, one of the 177 IMS acronyms — 159 different spellings of that column match, because it sometimes carries two work item names at once, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  4. 3509 IMS documents are of type CR IMS documents whose type column is exactly 'CR', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  5. 767 IMS documents are of type pCR IMS documents whose type column is exactly 'pCR', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  6. 175 IMS documents are of type CR pack IMS documents whose type column is exactly 'CR pack', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  7. 45 IMS documents are of type LS out IMS documents whose type column is exactly 'LS out', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  8. 40 IMS documents are of type WID revised IMS documents whose type column is exactly 'WID revised', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  9. 38 IMS documents are of type draftCR IMS documents whose type column is exactly 'draftCR', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  10. 19 IMS documents are of type WI exception request IMS documents whose type column is exactly 'WI exception request', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  11. 90 IMS documents are of type other IMS documents whose type column is exactly 'other', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  12. 61 IMS documents are of type discussion IMS documents whose type column is exactly 'discussion', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  13. 51 IMS documents are of type draft TR IMS documents whose type column is exactly 'draft TR', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  14. 32 IMS documents are of type draft TS IMS documents whose type column is exactly 'draft TS', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  15. 32 IMS documents are of type LS in IMS documents whose type column is exactly 'LS in', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  16. 30 IMS documents are of type WI status report IMS documents whose type column is exactly 'WI status report', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  17. 16 IMS documents are of type Work Plan IMS documents whose type column is exactly 'Work Plan', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05
  18. 1942 IMS documents carry the outcome word revised IMS documents whose status column is exactly 'revised', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05. The word is about that document and never about the work item
  19. 1736 IMS documents carry the outcome word agreed IMS documents whose status column is exactly 'agreed', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05. The word is about that document and never about the work item
  20. 379 IMS documents carry the outcome word approved IMS documents whose status column is exactly 'approved', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05. The word is about that document and never about the work item
  21. 297 IMS documents carry the outcome word postponed IMS documents whose status column is exactly 'postponed', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05. The word is about that document and never about the work item
  22. 189 IMS documents carry the outcome word withdrawn IMS documents whose status column is exactly 'withdrawn', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05. The word is about that document and never about the work item
  23. 173 IMS documents carry the outcome word noted IMS documents whose status column is exactly 'noted', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05. The word is about that document and never about the work item
  24. 68 IMS documents carry the outcome word merged IMS documents whose status column is exactly 'merged', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05. The word is about that document and never about the work item
  25. 51 IMS documents carry the outcome word not treated IMS documents whose status column is exactly 'not treated', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05. The word is about that document and never about the work item
  26. 35 IMS documents carry the outcome word rejected IMS documents whose status column is exactly 'rejected', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05. The word is about that document and never about the work item
  27. 24 IMS documents carry the outcome word endorsed IMS documents whose status column is exactly 'endorsed', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05. The word is about that document and never about the work item
  28. 20 IMS documents carry the outcome word not pursued IMS documents whose status column is exactly 'not pursued', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05. The word is about that document and never about the work item
  29. 19 IMS documents carry the outcome word available IMS documents whose status column is exactly 'available', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05. The word is about that document and never about the work item
  30. 6 IMS documents carry the outcome word partially approved IMS documents whose status column is exactly 'partially approved', asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05. The word is about that document and never about the work item
  31. 248 documents carry the work item IMSProtoc2 rows of `tdoc` whose work item column names IMSProtoc2 as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  32. 32 documents carry the work item IMSProtoc3 rows of `tdoc` whose work item column names IMSProtoc3 as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  33. 19 documents carry the work item IMSProtoc4 rows of `tdoc` whose work item column names IMSProtoc4 as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  34. 248 documents carry the work item IMSProtoc5 rows of `tdoc` whose work item column names IMSProtoc5 as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  35. 209 documents carry the work item IMSProtoc6 rows of `tdoc` whose work item column names IMSProtoc6 as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  36. 364 documents carry the work item IMSProtoc7 rows of `tdoc` whose work item column names IMSProtoc7 as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  37. 169 documents carry the work item IMSProtoc8 rows of `tdoc` whose work item column names IMSProtoc8 as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  38. 123 documents carry the work item IMSProtoc9 rows of `tdoc` whose work item column names IMSProtoc9 as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  39. 72 documents carry the work item IMSProtoc16 rows of `tdoc` whose work item column names IMSProtoc16 as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  40. 26 documents carry the work item IMSProtoc17 rows of `tdoc` whose work item column names IMSProtoc17 as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  41. 45 documents carry the work item IMSProtoc18 rows of `tdoc` whose work item column names IMSProtoc18 as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  42. 62 documents carry the work item IMSProtoc19 rows of `tdoc` whose work item column names IMSProtoc19 as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  43. 19 documents carry the work item IMSProtoc20-CT rows of `tdoc` whose work item column names IMSProtoc20-CT as a whole word, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05; the first and last meeting are those of its earliest and latest upload stamp
  44. 1680 IMS documents name 24.229 as the document they change IMS documents whose change-request column names 24.229, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05, of which 1675 are of type CR, CR pack or draftCR
  45. 339 of the 345 IMS work plan rows stand at 100% rows taken in by the acronym rule whose completion field reads 1, read 2026-08-05. The field is a percentage the work plan keeps by hand and it is not a status word
  46. 237 IMS documents carry no upload time IMS documents whose upload column is empty, asked of /var/www/whatthespec.net/data/database/api/api.sqlite on 2026-08-05, which is why the per-year counts do not add up to the total and why a count over a year and a count over the whole record answer different questions

Every source this course is built on

Check yourself

Answers appear when you pick one, with where they come from.

Q12.1 What kind of document makes up most of this subject's paperwork?

Q12.2 A change request carries the outcome word "agreed". What does that tell you?

Q12.3 How often has the IETF protocol alignment work been reopened?

Q12.4 The justification of IMSProtoc2 in Rel-8 and of IMSProtoc20-CT in Rel-20 are

Q12.5 339 of the 345 IMS work plan rows stand at 100%. What does that number mean?

This chapter was built from a source register generated 2026-08-05. A fresher build of the register may hold different numbers.