ISO/IEC JTC1/SC22 is the parent organization of WG9, the Ada working group. I attended the plenary meeting of SC22 in my role as convener of WG9.
I entered the meeting with several specific objectives for advancing the work of WG9. All of these objectives were achieved. They are represented by the following resolutions adopted by SC22 at the meeting:
ISO/IEC JTC 1/SC22 approves the registration of PDTR 15942, Information technology -- Programming languages -- Guidance for the use of the Ada programming language in high integrity systems (Type 3 Technical Report), Project JTC 1.22.15942.
ISO/IEC JTC 1/SC22 endorses the New Work Item Proposal for Ada: Conformity Assessment contained in document SC22 N2765, noting that five SC22 P-members (Canada, Germany, Japan, UK and USA) have agreed to participate in this work, and instructs its Secretariat to forward the NP for ISO/IEC JTC 1 approval.
ISO/IEC JTC 1/SC22, recognising that a Working Draft of the specification has been distributed to the SC22 Member Bodies for review and comment, authorises its Secretariat, subject to approval of the New Work Item Proposal by ISO/IEC JTC 1, to conduct concurrent CD registration and FCD approval ballots for Ada: Conformity Assessment.
ISO/IEC JTC 1/SC22 ... recommends that the following standards be withdrawn:
ISO/IEC JTC 1/SC22 approves the following appointments: ... Dr. Brian Wichmann (UK) as Project Editor for Project JTC 1.22.15942 -- Guidance for the use of the Ada programming language in high integrity systems (Type 3 Technical Report); ...
[All of the resolutions quoted above will undergo editorial revision and renumbering before they are formally published.]
I appreciate the efforts of the Heads of Delegation from Canada, Germany, Japan, UK and USA in arranging for their SC22 representatives to commit to active participation in the Conformity Assessment project.
Conveners are routinely required to provide schedules for all of their projects. I provided the following schedules:
Project | Name | CD (or PDTR) submitted for ballot | FDIS (or DTR) submitted for ballot | IS submitted for publication |
13813 & 13814 | Complex numbers | Done | Done | Done |
15291 | ASIS | Done | Done | 6/99 |
15942 | High Integrity | 10/98 | 10/99 | 10/2000 |
NP | Conformity Assessment | 11/98 | 4/99 | 7/99 |
WG20 continues its work on extended character sets. It is probably now appropriate for WG9 to pay additional attention to this issue. The C++ DIS and the planned C9X revision both incorporate extended character sets, even for use in identifiers. COBOL is planning to permit this usage. Fortran is not.
Currently TR 10176 provides Guidelines for Programming Language standards. It provides a list of "portable characters" based on IS 10646. (Another relevant document is CD 14652.) Characters are frequently added to 10646. WG20 wants the list of portable characters to be "easily updatable". Most of the conveners objected to this idea because the list becomes a moving target. The conveners want the list to be updated infrequently and version-controlled.
There are a variety of issues. One involves the use of extended character sets in identifiers. This can present problems for case-insensitive languages because the mapping between lowercase and uppercase can cause ambiguities. The conventional wisdom seems to be that the canonical form for identifiers in such languages should be the lowercase form.
Another issue is the use of extended digit characters in numeric literals. (For example, Kanji has its own set of characters to represent digits.)
[This is not a comprehensive report of the other working groups. Rather, a few items of interest have been selected.]
The APL projects are making little or no progress. Nevertheless, the projects were extended for another year. It seems clear that no more extensions will be granted.
ISO/IEC 8485:1989, APL, will be reaffirmed.
Several dependencies on the work of WG20 were cited. The primary concern of the Cobol group is the provision of a unchanging list of characters to be incorporated in any particular revision of the Cobol language. They do not want to refer to a document that continues to change.
WG5 has decided not to incorporate non-Latin characters in identifiers for Fortran 2000.
Interoperability with C is being incorporated into the Fortran standard.
ISO/IEC 1539-2:1994, Fortran -- Part 2: Varying Length Character Strings, will be revised. Concurrent CD registration and CD balloting was approved.
The terms of reference of WG11 were revised:
To study binding issues between programming languages and the abstract facilities with which they interface or interact, and to propose, develop and maintain ISO/IEC standards in this area.
ISO/IEC 10967-1:1994, Language Independent Arithmetic -- Part 1: Integer and Floating Point Arithmetic, will be reaffirmed.
The C9X revision will incorporate extended character sets for use in identifiers. It was reaffirmed that the C and C++ languages will be permitted to evolve independently with no requirements for mutual compatibility. For example, both languages will provide facilities for complex arithmetic, but those facilities will be different.
Denmark has been unsuccessful in convincing this working group to adopt a Danish proposal for the inclusion of standardized low-level I/O facilities. The convener of the C++ WG suggested that they might be interested in the work.
ISO/IEC 13210:1994, Test Methods for Measuring Conformance to POSIX, will be revised.
Two projects were withdrawn:
A new convener, Jim Oblinger, was appointed.
SC22 will also ask ITTF to reverse its recent decisions regarding the numbering of profile standards.
The JTC1 plenary has decided that WG20 shall organizationally remain in SC22, but will nevertheless be a part of the new Technical Direction "Cultural and Linguistic Adaptability". WG20 has nominated 3 representatives to the TD planning meeting in New York.
The ECMA standard on "Registration of European Cultural Conventions" is being adopted as ISO/IEC 15897. Although there are a lot of procedural and political issues outstanding, the project was assigned to WG20 for maintenance.
A word was added to the title of WG20: "Specification method for cultural conventions."
The FDIS for C++ was unanimously approved. They expect publication of the standard imminently. They plan to propose a New Work Item for the use of C++ in embedded and time-critical systems.
This WG has been disbanded upon the completion of their work.
The Java Study Group was continued.
As was anticipated, Sweden has decided to change its membership in SC22 from P to O. [Membership is separately maintained at every level of JTC1. I have already advised the Swedish delegation to WG9 that they are welcome to continue full participation in the work of WG9.]
The term of the current SC22 chair, Bob Follett, ends at next year's plenary. As he has served in this position for ten years, he does not plan to accept another term.
Mr. Isak Korn volunteered to serve as liaison officer from SC22 to SC7. Concern was expressed by some regarding a perceived overlap in the statements of scope of SC7 and SC22. No action was taken except that the SC22 Secretariat was asked to distribute the former and current statement of scope.