Ad-Hoc Meeting between IST/21/-/1/4 and IST/5/-/15
				     on
			OSE/Posix Conformance Testing


	The meeting took place on 10-Mar-1994 at Hampden House, London,
	beginning at 10:30.  Those present were:

		Dave Rayner	IST/21/-/1/4
		Ian Valentine	      "
		David Cannon	IST/5/-/15
		Colin O'Driscoll     "

	The morning was spent describing the area of work of each of the
	two panels;  CO'D and DC described the structure of the Posix
	standard and the method of working of the IEEE PASC working
	groups and of ISO/IEC WG15 as a basis for outlining the Posix
	Conformance Testing drafts and what they were intended to cover.
	The shortcomings of this work were described, and this lead to a
	description of the UK draft NP on OSE Conformance Testing
	Methodology which the IST/5/-/15 Panel had drafted and taken
	through WG15 for SC review and comment.

	In exchange, DR and IV described the development of the IS 9646
	series of standards, determining OSI Conformance.  They pointed
	out that the scope of that work was deliberately restricted to
	OSI standards.	An outline of the work is:

		Std			1st Edn 2nd Edn

		9646-1
		General Concepts &	1991	1994
		Terminology

		9646-2
		ATS Specification	1991	1994

		9646-3
		TTCN			1993
				(Amd-1 in 1994)
					 (Revision in 1995?)

		9646-4
		Test Realisation	1991	1994

		9646-5
		Test Labs & Client	1991	1994
		Report Formats

		9646-6		       (1994)
		Protocol Profile
		Test Specification*

		9646-7			DIS
		Implemenation
		Conformance statement

	(* - just the Application and Transport layer blocks of the OSI
	     model, not the set of 7 discreet layers)

	The second edition, most of which is targeted for 1994, adds
	Profiling and multi-party functionality.  In doing so some
	tracts of text are moved from one section of the 9646 standard
	to others, in order to retain the overall structure.

	Some discussion followed this exchange of background on whether
	Posix addressed the issue of application portability.  This
	issue crops up again and again; although the Posix standards do
	make a	number of assertions about 'strictly conforming'
	applications, etc, the Posix standards describe an application
	platform, and how to implement a conforming version of that, not
	how to implement an application which will execute on the
	platform.

	There was also a discussion on what OSI APIs fitted under the
	Posix umbrella and whether this was an appropriate place for
	further work.  The IEEE mechanism for new projects was outlined,
	but it was suggested that SC22 was possibly a more suitable
	target, being titled "Programming Languages, their Environments
	and System Software Interfaces", unless the proposed API was
	closely related to Posix.


	DR pointed out that although the SC21-sponsored NP on
	Conformance Testing Methodology had failed its JTC1 ballot, the
	work was needed for the emerging ODP standard, amongst others.
	From IST/5/-/15's point of view, a work item which addressed the
	concerns raised in WG15 N423 "NP on Guide for conformance
	testing methodology (UK)" would be welcomed - the problem with
	the original SC21 NP was that it missed this, and the existing
	Posix TM work, completely.


	There is an upcoming SWG-CA (Special Working Group - Conformance
	Assessment) meeting in London from 20-22 June, and subsequently
	in Korea in October.  A call for contributions is currently out.
	DR was keen to take advantage of this opportunity, and it would
	seem that IST/5/-/15 should make some effort to get the N423
	document tabled for this group to consider.  The SWG-CA meeting
	in London is open to all-comers:  IST/5/-/15 might like to send
	some attendees.

Action: IST/5/-/15 to consider nominating Conformance people to attend
	SWG-CA in London, 20-22 June, for the purpose of explaining the
	WG15 N423 "NP on Guide for conformance testing methodology (UK)"
	proposal and inviting comment.

Action: DC to note the upcoming SWG-CA meeting in London, 20-22 June, at
	the May meeting of WG15.

	DR noted that at its last meeting EWOS approved a new work item
	to analyse the existing Conformance Testing Methodology (CTM)
	work, identify gaps, but NOT develop testing methodology.  EWOS
	does not see itself as a TM group.  At the same meeting EWOS
	authored a number of requests for liaisons, including one to
	WG15, specifically to cover Profile and CTM issues.  This will
	probably be considered at the WG15 May meeting.


	After lunch the group discussed technical issues in a number of
	papers, including one authored by Gerry Powell describing
	differences in terminology between OSI and Posix CTM work, and
	one from the NCC (?), thought to have been superceded,
	describing the pragmatics of Posix and OSI test suites.  CO'D
	offered to check on the availability of a more recent version.


	The meeting closed at 15:00