This article describes how the original GSM Specifications were drafted and published when they were the sole property of ETSI, long before the creation of 3GPP.
Work started on a harmonized European cellular technology in the early 1980s under the auspices of CEPT, carried out in an ad hoc subgroup called Groupe Spécial Mobile - GSM for short. An overview of the early history of GSM can be found in Hillebrand (ed) 
ETSI was created in 1988 and gradually all standardization work was transferred from CEPT to ETSI, normally retaining the existing committee structure. CEPT GSM became ETSI Technical Committee (TC) GSM in 1990, and in 1991 the TC was renamed "Special Mobile Group", SMG. The original CEPT GSM had had a dedicated technical secretariat or "Permanent Nucleus", based at the CEPT premises in Paris, and this too eventually transferred to ETSI headquarters in Sophia Antipolis. The Permanent Nucleus was thereafter known as Project Team (PT) 12 or, informally, as PT SMG. ETSI’s Project Teams were later renamed Special Task Forces so PT12 became STF12.
The Permanent Nucleus and PT12
Permanent and Temporary Documents
On the move of work from CEPT to ETSI, the paper archives of CEPT were also transferred. These, though far from complete, are still kept safely at ETSI, together with the archives of ETSI TCs GSM and SMG. In the mid 2000s all the paper archives were scanned to PDF with a view to making them more accessible.
ETSI worked with master documents on paper until the mid-1990s, when a gradual move was started to all-electronic working. The document archive of TC SMG is therefore much more up to date from around 1997 onwards.
It was decided early on after the transfer of work to ETSI that the pan-European digital cellular standards could not be developed to completion in one go, and that a phased approach should be adopted. Thus a basic telephony service could be standardized as a "Phase 1", which would be followed by a Phase 2, and so on, each new Phase adding more and more functionality. A set of GSM documents (still essentially in the form of old CEPT Recommendations, though published over a period of a year or two as a mixture of "GSM Technical Specifications" (GTS), ETSs and ETRs) was published to represent the Phase 1 offering.
Subsequently - logically - an expanded set of publications formed the Phase 2 system.
At this point (c.1996) it was realized that, in the fullness of time, the second generation cellular system (GSM) would one day give way to a more performant 3rd generation system. (The ITU was referring to this technology-to-be as the somewhat unwieldy FLMTS - Future Land Mobile Telecommunication System -, while in ETSI this became known, perhaps slightly presumptuously, as UMTS - Universal Mobile Telecommunication System.) For this reason, the next release of the GSM specification set was termed not Phase 3 but "Phase 2+", and this nomenclature was retained for subsequent specification sets, which became known as Releases, finalized more or less annually, as R97, R98 and R99, with the original Phase 2+ being referred to R96.
TC GSM/SMG had adopted a three-field version number system, but ETSI’s publication scheme dealt with "editions". ETSI had also devised a system of making minor additions to a published ETS or ETR in the form of "Addenda", and also of making minor corrections to errors discovered after publication in the form of "Corrigenda". Several versions of GSM specs are published as addenda or corrigenda to original ETSs. The GSM publications were unusual in that, because of the phased approach, the output documents never became fully stable, but were continually revisited, with updates being made to some of the more complex documents at nearly every plenary meeting. This posed a problem for ETSI, whose publication mechanism mirrored the traditional, lengthy bureaucratic process long used by CEN and CENELEC. Because of the national public inquiry, revision, and national vote process, an ETS took a minimum of eight months between TC approval and final publication. By the time a GSM ETS was published, the TC might have already revised it three times, making the ETS already out of date! This incompatibility in working methods was a source of some friction between the "traditional" ETSI community and the GSM community in the early days. Publication of a GTS was quick and dirty (insofar as there were no public inquiry and vote phases), but satisfied the GSM community; on the other hand, the European Commission would not except such unscrutinized publications for use as the technical basis for type approval testing. These problems were eventually overcome more by market forces as the GSM service rapidly became very popular with its users, rather than by resolution within ETSI.
Where, as was the case for most GSM publications, the final document type was to be an ETS, the course of events might be something like this ...
At plenary meeting number N, TC SMG would approve (say) version 3.4.0 as being stable enough to issue as an ETS. This would be transferred to the ETSI Secretariat, which would perform a little cosmetic clean up and issue it for national public inquiry. This version of the document was know as the "first draft" of the ETS. During the course of the public inquiry, TC SMG would continue to work on the document and might hold as many as two more plenary meetings (N+1 and N+2) during the public inquiry; and by the time the comments resulting from the public inquiry were received by the Secretariat, collated, and passed to TC SMG, the latest version of the document might have progressed to (say) 3.6.0. TC SMG then had the potentially difficult job of incorporating remedies to those national comments on the earlier version into the now current version of their document, which would then be raised to version 3.7.0 and sent to the Secretariat which would prepare it for the national vote as a "final draft" of the ETS. At the end of the vote, assuming a positive result (no GSM ETS ever failed the vote stage!) the Secretariat would publish the ETS as edition 1. Meanwhile TC SMG might have further modified its specification to version 3.8.0.
This somewhat confusing situation was somewhat alleviated after Phase 1 by the ETSI-issued ETSs also bearing the three-field SMG version number on the page headers, leaving the reader in no doubt which version of the GSM Spec he was reading. The situation was further improved some years later, when ETSI’s publication regime was brought more in line with CEN’s and CENELC’s, with the published standards changing from "ETS" to "EN" (Europäische Norm) and edition numbers being dropped in favour of a three-field version number based upon that invented for the GSM Recommendations. It is important to note that the ETSI version numbering system was not identical to that devised for GSM because, whilst the GSM version would be incremented each time the document was edited, however small the change, the ETSI ETS retained the same version (a) when it was converted to the ETSI "first draft" ETS for public inquiry, (b) when it was converted from the SMG-updated version incorporating the latest modifications including the public inquiry comments to the "final draft" ETS for national vote, and (c) when it was converted from the "final draft" to the document of the final, published ETS. Thus, in the extreme case where TC SMG made no changes to a document during the public inquiry phase and where no comments were received during the public inquiry (European industry rapidly learnt not to rock the GSM boat and to refrain from any but absolutely vital comments during the public inquiry!), the same version number could appear on four different instances of the document:
- that tacitly approved by TC SMG after incorporation of the latest approved CRs by PT12,
- the "sanitized" editorial modification sent out by the ETSI Secretariat as the "first draft" ETS  for public inquiry,
- the editorially revised version sent as the "final draft ETS" for national vote, and finally
- the again editorially revised version published as the definitive ETS.
Because of this, a certain amount of care has to be taken when discussing a particular version of a GSM Specification. There could be up to four instances of a GSM document identified by the same version number.
The situation was improved somewhat by adopting the slightly streamlined ETSI process originally known as the Unique  Approval Procedure (UAP), later renamed the One-Step Approval Procedure (OAP), whereby the national public inquiry and vote phases were combined, the EN being published in its original form, with any comments received being fed back to the responsible Technical Committee with a view to issuing a subsequent version.
Later still, use was increasingly made of the even more efficient publication process which involved neither national public inquiry nor vote, but publication immediately after approval by the TC; this process was not permitted for ENs but was restricted to a different deliverable type, the Technical Specification, originally seen as having lesser status than ENs by virtue of this much more restricted publication cycle, but much more appropriate to TC SMG’s working methods. Purely informative material was published in Technical Reports, similarly free from burdensome approval procedures.
By the late 1990s, this regime was used for all GSM publications except those intended for regulatory type approval testing (the 13.-series specs), and on formation of 3GPP in 1998, the same approach was taken, with all 3GPP normative output being of type Technical Specification and informative output being of type Technical Report.
Transfer to 3GPP, closure of ETSI TC SMG
The new collaborative venture 3GPP was formed in 1998 and held its first plenary meetings in December of that year. Initially, change requests were approved twice: by 3GPP and by TC SMG, but gradually, the non-radio GSM specs were transferred to 3GPP ownership and this dual approval vanished.
Finally, the remaining GSM specs relating to radio were transferred to a newly-created 3GPP Technical Specification Group "GERAN" in mid-2000, and TC SMG was closed. Only the EU-regulatory standards were retained within ETSI, and these were transferred from SMG to a newly created Technical Committee, "Mobile Standards Group" (MSG).
ETSI has issued several CDs and DVD sets relating to the development of the GSM system:
- Reports of all TC GSM and SMG plenary meetings, plus all available contribution documents to SMG plenaries.
- Development of the SIM card: All available contributions to the GSM SIMEG and later SMG9 working groups.
- A Technical History of GSM: All available contributions to TC GSM and TC SMG meetings, both plenary and working groups, plus all available draft GSM specifications and published versions up to the end of Release 99.
These are available for purchase to ETSI member organizations.