NOTES
The 32nd meeting of ISO/IEC JTC1/SC22/WG9 (Ada) will be held Monday, 2 June 1997, in the Mortimer Room of the Tower Thistle Hotel, London, England, United Kingdom in conjunction with the Ada-Europe conference {http://www.sis.port.ac.uk/adauk/Ada-Europe/Conference/1997/}. The meeting will begin at 9:00 am and adjournment is anticipated for 4:00 pm. We will recess for a one-hour lunch at approximately 11:40. Each individual will make his/her own arrangements for lunch.
Contact Information for the hotel follows:
Tower Thistle Hotel, St. Katharine's Way, London, E1 9LD Tel: (0171) 481 2575 Fax: (0171) 488 4106ACTION 31-1 (Moore): Identify an individual or a national body willing to assume the duties of recording secretary. Closed: In addition to serving as WG9's Webmaster, Clyde Roby has agreed to serve as Secretary at meeting #32.
The (currently unapproved) minutes of the Philadelphia meeting are available as document N323 {http://www.open-std.org/jtc1/sc22/wg9/documents.html#N323}. The resolutions approved at that meeting are available as document N324 {http://www.open-std.org/jtc1/sc22/wg9/documents.html#N324}.
After publication of N323, Kiyoshi Ishihata offered corrections. All have been incorporated into the resolution below except for a single comment:
Convener recommends omitting this correction because the tasked individuals understood the action item and will appropriately respond later in this meeting, closing the action item.
RESOLUTION: The Minutes of Meeting #31 (N323) are approved with the following corrections:
Convener's recent report to SC22 can be found at {http://www.open-std.org/jtc1/sc22/wg9/documents.html#N328}.
James Moore has now been appointed convener by SC22.
WG9 Web page is located at {http://www.open-std.org/jtc1/sc22/wg9/}. Clyde Roby is Webmaster.
Since the last meeting, some progress has been made in resolving the copyright issue with ISO. For a one-year trial period, documents below the DIS level are excused from the copyright requirement. A Special Order on this subject is scheduled later in the agenda.
JTC1 and SC22 are rapidly moving toward distributing documents only in electronic form. A Special Order on this subject is scheduled later in the agenda.
The normal progression of documents is now as follows:
The DIS designation is now only used for fast-track ballots.
From Robert Follett, Chair of SC22 -- Development of a standard for a language binding requires knowledge and expertise in two areas - the standard for the technical subject of the binding and the standard for the programming language involved. Thus, for example, the development of the GKS Fortran binding required expertise in both GKS and Fortran. It has generally been the case in the past that the binding is best produced by experts in the technical subject of the binding rather than by programming language experts. As a result, the language bindings developed in SC 22 have involved bindings to technical areas within SC 22, e.g., POSIX or PCTE, but not to technical areas in other SCs, e.g., graphics bindings in SC24. SC 22 believes this division of work is appropriate. There may be times when a language binding would be produced by the language experts but such cases should be assigned on a case by case basis and are not the general rule.
As planned at the last meeting, the ASIS Working Draft was balloted by national delegation and approved. Upon approval of the comment disposition report, later in the agenda, a Committee Draft will be forwarded to JTC1.
Convener plans to attend the forthcoming SC22 plenary in Ottawa, 18-22 August 1997. Rapporteur Group chairs are reminded of the following action item:
ACTION 31-3 (Rapporteur Group Chairs): WG9 encourages its Rapporteur Groups to schedule their meetings in conjunction with the SC22 plenary in Ottawa, 18-22 August 1997. Each RG chair should correspond with Steve Michell regarding their intentions. Open: RG Chairs are still encouraged to schedule meetings in conjunction with the SC22 plenary meeting.
Attendance by the following delegations and individuals is currently anticipated:
RESOLUTION: Meetings #33 and #34 of ISO/IEC JTC1/SC22/WG9 are scheduled as follows:
Rapporteur Groups are encouraged to meet in Ottawa, Canada in conjunction with the SC22 Plenary meeting, 18-22 August 1997. The point of contact for arranging meetings is Steve Michell.
JTC1 and SC22 are moving aggressively to eliminate paper distribution of documents and to move toward electronic distribution methods, even for balloting. On the other hand, ISO seeks to impose its copyright upon all products of the standardization product. A one-year trial period has been approved during which the requirement for ISO copyright upon documents belong the level of DIS has been suspended.
Additional information regarding these changes can be found at a Web site dedicated to this purpose, {http://www.open-std.org/JTC1/IMPIT/} particularly {http://www.open-std.org/JTC1/IMPIT/open/j1n4558/j1n4558.htm}.
You should be prepared to read documents in any of the following formats:
WG9 will continue to use its Web site to make documents available. At some point, we will set up a password-protected Web site also. We are capable of doing this whenever it is desired.
For informal documents, HTML and plain ASCII text are preferred methods of distribution. RTF may be preferable to either Word or WordPerfect. Word documents should be designed so that understandability is not hampered by changing page size between 8-1/2 x 11 and A4. The use of the default "Normal" style for text should be avoided.
The Type 2 Technical report was published 14 November 1996.
RESOLUTION: Having completed their chartered mission, the Real-Time Rapporteur Group is discharged, as of Meeting #31, with the grateful appreciation of ISO/IEC JTC1/SC22/WG9. The RRG has made important contributions to Ada technology through the drafting of Type 2 Technical Report 11735, describing real-time extensions suitable for use with Ada.
ACTION 31-12 (Moore): Again query the national bodies regarding their interest in Ada bindings to SQL and ODBC. Based upon their responses, make recommendations concerning the standardization of such bindings. Closed: No national bodies indicated they were willing to actively support such an effort; three replied that they could not support an effort.
RESOLUTION: Having completed their chartered mission, the SQL Rapporteur Group is discharged with the grateful appreciation of ISO/IEC JTC1/SC22/WG9. The SRG has made important contributions to Ada technology through the drafting of the SAMeDL standard, IS 12227.
ADMINISTRATIVE ACTION: Subject to approval by SC22, the convener appoints Erhard Ploedereder and Bob Duff as Project Editors to perform the continuing responsibilities associated with IS 8652:1995.
RESOLUTION: ISO/IEC JTC1/SC22/WG9 expresses its grateful appreciation to Chris Anderson and Tucker Taft for their accomplishments as Project Editors of IS 8652:1995.
ACTION 31-9 (Moore): Forward copies of defect reports from other language working groups to the Chair of the ARG. Closed. This was done.
ACTION 31-8 (Moore): After resolution of the ISO copyright issue, determine a schedule for preparation of a technical corrigenda to the Ada standard, ISO/IEC 8652. Closed. To be discussed at Meeting 32.
ACTION 31-2 (Moore): Develop a plan for processing Ada Issues that provides for public access to the Issues during a discussion period while appropriately conforming with ISO copyright rules regarding the preparation of Technical Corrigenda. Closed. To be discussed at Meeting 32.
DISCUSSION
Received from Erhard Ploedereder, Additional Editing by Bob Duff: All referenced AIs can be found on the WWW pages of WG9 by following the Rapporteur group, then the ARG, and then the {Ada Issues} links. On this page, the referenced AIs are found under the "arg-approved" category. (Note that this category also includes some other AIs that the ARG will bring to WG9 at the fall meeting.
RESOLUTION: To approve the following AIs which have been previously approved by the ARG:
Note: Some of the AIs are still undergoing editorial review by the ARG, which will be concluded before the WG9 meeting. In case of any substantive changes, the AIs will be selectively withdrawn from the motion. [Two changes of titles have been incorporated into the preceding motion -- jwm]
ACTION 31-6 (Moore): For meeting #32, designate an agenda item for discussion of ARG direction and an agenda item for consideration of Ada Issues related to C interfaces. Closed.
ACTION 31-7 (Ploedereder): Prior to meeting #32, prepare an outline of issues related to C interfaces so that delegates may prepare themselves for a discussion of these issues. Closed.
Received from Erhard Ploedereder: Below is advance material for a briefing on problems with C interfacing, action item # 31-7.
There are several issues in this area:
There is a fourth issue in the making, dealing with bindings to Java Interfaces.
The global question is: how progressive and how (in)compatible can changes be to make Ada more easy to interface with other languages ?
At issue is the parameter passing convention for records passed to a subprogram with convention C (AI-131). The RM contains Implementation Advice to use by-reference passing in this case. B.3(69). However, the C standard passes structs by-value. While this Implementation Advice models the C programming idiom to often pass structs by reference, i.e., as t* rather than t, and hence is a convenience in many cases, it also makes it impossible to interface to a C routine that does not adhere to this idiom.
The possible actions are:
The ARG is split on the issue. In Montreux, 2/2a was the preferred alternative, in Vermont opinion swayed to 1, in Henley it received a 2:3:5 vote.
At issue are the language rules regarding the convention of dispatching operations. Presently, the rules specify that their calling convention is Ada, unless specifically overridden. On the other hand, there is the rule that overriding subprogram declarations must have the convention of the overridden subprogram. (With friendly reading of the RM, inherited subprograms inherit the convention of their parent, rather than being always of convention Ada.)
If tagged types are interfaced to OOP types in another language, e.g. C++, the imported operations will have to be of convention C++. This leads to two problems:
The ARG therefore proposes new rules, which are reasonably upward-compatible and much more friendly:
DISCUSSION
ACTION 31-4 (Heads of National Delegations): Determine a point of contact for the imminent letter ballot on ASIS. Closed.
ACTION 31-5 (Moore): Conduct the letter ballot on ASIS. Closed. The letter ballot on the ASIS Working Draft was completed with six nations (Germany, Japan, Russia, Switzerland, UK, USA) voting affirmatively and one nation (Sweden) abstaining. No negative ballots were cast.
In accordance with the plan approved at our meeting of December 1996, the ASIS Project Editor is to prepare a revised Working Draft in accordance with the comments received during the balloting and prepare a Comment Disposition Report for consideration at Meeting #32. Approval of the Comment Disposition Report at Meeting #32 will authorize the Convener to forward the revised Working Draft to SC22 for CD registration.
On 26 April 1997, convener received a report from the chair of the ASIS RG, reading in part:
The ASISRG met from 25-26 March 1997 in Phoenix, Arizona and developed a general strategy for the disposition of comments. All comments have been categorized as either Editorial Comments or Technical Issues. Ninety (90) Editorial Comments and fifty-two (52) Technical Issues were received as a result of the WG9 balloting process of ASIS.
The disposition of both the Editorial Comments and Technical Issues is provided via the ASIS Home Page at the {http://www.acm.org/sigada/WG/asiswg} URL. An overall Summary of the Editorial Comments is provided by following the link to ISO Working Draft and then to {Editorial Comments}. Each Editorial Comment with its disposition can be reviewed there by following associated links.
Technical Issues received as part of the WG9 balloting process were added to the existing ASIS Issues Summary, starting with Issue #023. An overall Summary of the Technical Issues is provided by following the link to ISO Working Draft and then to {ASIS Issues}. Each Technical Issue with its disposition can be reviewed there by following associated links. There is also a link to additional email discussion of most issues.
The ASIS Working Draft has been revised to incorporate approved comments. It is denoted ASIS Version 2.0.M. This version contains all ASISRG Approved ("A") / Approved as Modified ("M") Editorial Comments and Technical Comments except:
Also the examples in Annex B have NOT been compiled to this new version. The examples will be compiled and executed using an updated ASIS implementation and actual results of the execution will be provided. Outstanding comments and examples will be incorporated in ASIS Version 2.0.N, to be distributed to WG9 at their meeting on 2 June 1997. ASIS Version 2.0.M is available on the ASIS Home Page following the link to ISO Working Draft. It is provided in both a Microsoft Word (.doc) file and as a Postscript (.ps) file. The compilable portion of the specification is provided in ASCII.
The ASISRG thanks WG9 for the many excellent comments you have provided. Your input has resulted in a significantly improved ASIS interface to the benefit of the entire Ada community.
RESOLUTION: The disposition of comments on the ASIS Working Draft ballot is approved by WG9 and the Project Editor of project 15291 is authorized to forward Version 2.0.N of the draft to the SC22 secretariat for CD registration.
In February 6/7th 1997, the HRG meet, the main topic of which was to resolve the remaining issues relating to the structure and overview of the proposed Guidelines (see the HRG Charter, already agreed by WG9).
The principle change over the previous outlines of the Guidelines was not to attempt to have 'levels' as in IEC 1508 (draft) or DO-178B, but to use the analysis/ verification methods as the main aspect which determines which Ada language features are to be recommended. This method is explained in the paper 'High Integrity Ada' which is in a following e-mail message (as a PostScript document).
The HRG agreed to submit to WG9 a proposal for the New Work Item for the Guidelines (to be a type 3 Technical Report). This NWI proposal is in a following e-mail message as an HTML file.
The HRG also agreed to ask WG9 to formally make Gilles Motet a member of the HRG. He provided excellent support of the meeting which was a great success.
If any WG9 member requires this information as a printed set of documents, please e-mail your postal address to me.
can be found at: {http://www.open-std.org/jtc1/sc22/wg9/HRG-NWI.html}
JTC1 now places additional requirements upon the approval of new work items. The following material from the Annex H RG chair addresses these requirements:
This document is a copy of ISO/IEC JTC1 N4477, to which has been added, in italics, the relevant information on the proposed HRG Guidelines.
The existing JTC 1 Directives (cl 6.2.1.3) require the commitment of five National bodies; in addition the criteria (proforma) defined in this paper shall also be satisfied.
The HRG has active support from members from the following countries: Canada, France, Netherlands, UK, and USA. We are hoping for activity participation from Germany.
My working assumption was that the NP would be initiated by SC22, but I assume the UK could also initiate this.
Ada 95 is specifically designed to handle high integrity applications. Large high integrity applications almost always involve more than one country, and therefore international guidance is required which system certifiers can use. The system certifiers are often not Ada experts and therefore need guidance in which they can have confidence.
Ada 95 is not currently being used for high integrity application, but this will change shortly. Hence it is important that the Guidelines are produced with a short period to satisfy the future requirement.
For technical details for the proposed Guidelines themselves, with a list of participating members, see the paper High Integrity Ada.
I have previously circulated the old proforma, as per the ASIS NWI. The new one is part of this document.
I hope we have established in the HRG that we have the resources and expertise.
The following proforma is proposed for providing evaluation criteria for NP acceptance.
The procedure for NB determination of a NP ballot as in the existing procedure at the NB level. There is now with this proposal proforma additional information to aid NB's in their determination. The NB's when responding to the NP shall, where possible, relate any comments to the criteria assessments made by the proposer on the Proforma.
That which identifies market place relevance in terms of what problem is being solved and or need being addressed.
There is an international market in high integrity software within aircraft, defence and railways systems (for instance). Many of these systems are being written in Ada. International Guidelines will directly support this market.
When submitting a NP, the proposer shall identify the nature of the Market Requirement, assessing the extent to which it is essential, desirable or merely supportive of some other project.
It is frequently the case that the prime contractor for a system subcontracts the high integrity software to another company, even in another country. If safety is involved, such as with aircraft, there will be a certification process involving a regulatory. Hence three parties are involved here, but with large projects many more parties need to work together in a framework which will produce a satisfactory product. For using Ada 95, it is thought that Guidelines are virtually essential.
If a Regulatory requirement is deemed to exist - e.g. for an area of public concern, e.g. Information Security, Data protection, potentially leading to regulatory/public interest action based on the use of this voluntary international standard - the proposer shall identify this here.
Almost all high integrity software is covered by some regulation. However, the actual regulations vary with the application domain (safety, security, etc) and industrial sector (defence, civil avionics, railways, etc). The proposed Guidelines are being designed to support all these regulations.
Aspects of the relationship of this NP to other areas of standardization work shall be identified in this section.
The Guidelines are totally dependent upon the Ada standard, ISO 8652.
If this NP is concerned with completing or maintaining existing standards, those concerned shall be identified here.
Not relevant.
Groups, bodies, or fora external to JTC1 to which a commitment has been made by JTC for cooperation and or collaboration on this NP shall be identified here.
Not relevant.
If other activities creating standards or specifications in this topic area are known to exist or be planned, and which might be available to JTC1 as PAS, they shall be identified here.
Not relevant.
The proposer shall indicate here an assessment of the extent to which the proposed standard is supported by current technology.
Indicate here the extent to which the technology is reasonably stable and ripe for standardization.
High integrity applications have been developed successfully in Ada 83, the previous Ada standard. This proposal will exploit the specific additions in the current Ada standard.
If the NP is anticipatory in nature on expected or forecasted need, this shall be indicated here.
The only anticipation is that of high integrity applications moving from Ada 83 to the current Ada standard.
If the NP relates to the creation of supportive reference models or tools, this shall be indicated here.
It is expected that software suppliers will provides tools to assist in the application of the Guidelines.
Any other aspects of background information justifying this NP shall be indicated here.
See the paper `High Integrity Ada'.
As of May 16, national bodies committing some level of active support are: Canada, France, Japan, Netherlands, UK and US.
The chair of the Annex H RG has provided a supplementary paper, High Integrity Ada, which can be found at: {http://www.open-std.org/jtc1/sc22/wg9/HRG-High-Integrity-Ada.html}, {as a PostScript file (57K bytes)}, and {as a text file (33K bytes)}.
ACTION 31-10 (Moore): Contact Don Sando to determine the state of DIS 13813. Closed: Mathis did this.
ACTION 31-11 (Mathis): Submit DIS 13814 for JTC1 ballot after ensuring that there are no dependencies upon the uncompleted DIS 13813. Closed: Moore submitted both items for DIS ballot.
RESOLUTION: Having completed their chartered mission, the Numerics Rapporteur Group is discharged with the grateful appreciation of ISO/IEC JTC1/SC22/WG9. The NRG has made important contributions to Ada technology through: (1) the drafting of supplementary arithmetic standards 11430 and 11729; (2) the drafting of the Numerics Annex of the Ada 95 standard; and (3) the drafting of two DIS documents for Complex Arithmetic. The NRG's achievements have been widely recognized including recognition as recipients of an Ada Distinguished Service Award, presented by SIGAda.
The indicated individuals will continue to be listed as Project Editors of the existing standards and the two new drafts. As many of you know, funding problems endanger the active support of these editors. So further changes may become necessary.
ACTION 31-13 (Moore): Examine the report of the SC22 Secretariat to summarize the status of LIS projects from WG11. Closed: A summary can be found at: {http://www.open-std.org/jtc1/sc22/WG11/}.
This Working Group has expressed a desire to work closely with other language WGs to specify inter-language calling conventions.
From Bob Mathis, Convener, SC22 Java Study Group (JSG) -- Sun decided to apply for PAS (Publicly Available Specification) submitter status and use that approach for initial standardization of Java. The JTC1 ballot on this application closes July 14 so the results at not yet known. There has been some discussion on this topic on the JSG mailing list. There have been more formal public comments associated with the US voting process.
ECMA TC-39 has completed its initial work in developing a standard based on JavaScript and JScript. I expect it to be approved by the ECMA General Assembly at the end of June and submitted for ISO Fast Track processing very soon after that. Anyone interested in reviewing that standard should contact me directly.
The JSG will hold its next meeting, June 30 - July 1, 1997, at BSI in London, UK.
More information can be obtained: about the Java Study Group from {http://www.open-std.org/jtc1/sc22/JSG}, about the US consideration of Sun's application from {http://www.jtc1tag.org}, and about the ECMA TC-39 from {http://www.ecma.ch}.
Participation in the work of JSG is welcomed; information about the mailing list is on the Web page.
Other questions can be sent to me directly: Bob Mathis {bob@pithecanthropus.com}.
IEEE Std 990 (PDL/Ada) will be administratively withdrawn this year because there is no interest in updating it to the level of Ada 95 and object-oriented technology.
RESOLUTION: Having completed their chartered mission, the Uniformity Rapporteur Group is discharged, as of Meeting #30, with the grateful appreciation of ISO/IEC JTC1/SC22/WG9. The URG has made important contributions to Ada technology through the promotion of uniform implementation of Ada language features.
ACTION 30-1 (Landwehr): Landwehr will take from the URG chair the list of existing UIs thought to be applicable to Ada 95 and prepare an informal document for discussion via email. Administratively closed.
RESOLUTION: WG9 gratefully recognizes and appreciates the fine work of Clyde Roby in developing and maintaining the WG9 Web Page.
The following Rapporteur Group Chairs are reappointed: Erhard Ploedereder (ARG), Currie Colket (ASIS RG), and Brian Wichmann (Annex H RG).
Subject to approval by SC22, the convener appoints Andreas Koeller as Project Editor to perform the continuing responsibilities associated with IS 12227:1995.
Adjournment at 4:00 pm is anticipated.