Draft Minutes of July 1999 Meeting (Austin/3)
1. Opening & Introductions
Welcome to the 3rd Joint PASC/TOG/WG15 Revision Group meeting. Andrew Josey called the meeting to order at 9.08am Tuesday July 20, at the Best Western Europa Hotel, Montreal, Quebec.
1.1 Introductions
1.1.1 Attendance list
Name |
Position |
|
Andrew Josey |
Chair, TOG |
ajosey@opengroup.org |
Jim Zepeda |
HP, TOG |
jzepeda@cup.hp.com |
Frank Prindle |
DoD, TOG/PASC |
prindle@voicenet.com |
Lee Damico |
Sun, TOG |
lee.damico@eng.sun.com |
Cathy Hughes |
Editor, TOG |
c.hughes@opengroup.org |
Don Cragun |
Sun, TOG/PASC |
dwc@eng.sun.com |
Nick Stoughton |
Secretary, PASC/WG15/TOG |
nick@usenix.org |
Andrew Gollan |
Sun, TOG (XNET) |
adjg@eng.sun.com |
Ulrich Drepper |
Cygnus, Linux Community |
ulrich@cygnus.com |
Peter Anvin |
Transmeta, Linux Community |
hpa@transmeta.com |
Michael Gonzalez Harbour |
Universidad De Cantabria,PASC |
mgh@ctr.unican.es |
Ted Baker |
Florida State University, PASC |
baker@cs.fsu.edu |
Mark Brown |
IBM, TOG |
mbrown@austin.ibm.com |
Curtis Royster |
DoD/DISA, PASC/TOG |
roysterc@acq.osd.mil |
Yasushi Takahata |
Mitsubishi/WG15 |
takahata@isl.melco.jp |
Keld Simonsen |
DKUUG, WG15 |
keld@dkuug.dk |
Lowell Johnson |
IEEE, PASC |
l.johnson@computer.org |
Finnbarr Murphy |
Compaq, TOG |
fpm@zk3.dec.com |
Joanna Farley |
SCO, TOG |
joannaf@sco.com |
1.2 Approve Agenda
Andrew outlined the overall purpose of the meeting, giving some of the choices that we need to address.
The agenda (AUSTIN/27) was approved without addition.
1.3 Approval of Minutes
The minutes of the previous meeting (AUSTIN/21) were approved without modification.
1.4 Minutes of Previous meeting (AUSTIN/7)
The previous minutes were approved as distributed.
1.4.1 Matters Arising - Action Item Review
A9809-09 MB to add columns to AUSTIN/6 for what .13, FIPS, & SUS require to be able to control.
OPEN
A9809-09 MB to add columns to AUSTIN/6 for what .13, FIPS, & SUS require to be able to control.
CLOSED (OBE)
A9903-01 AJ to obtain a copy of .1n (corrigenda) for reference CD.
CLOSED
A9903-02 NS and AJ to prepare the other three PARs for the PASC-PMC in case they are required.
CLOSED
A9903-03 AJ to ensure that all networking bodies (e.g. XoTGNET and PASC-Distributed Systems) realize that they are explicitly invited to Austin Group meetings.
CLOSED
A9903-04 AJ to provide KS with some examples of how a function (e.g. fseek) might look in time for the next WG14 meeting in London, WB 21 June.
LEAVE OPEN - change to October 1999 meeting
A9903-05 Keld Simonsen to supply a list of overlapping functions between XSH and ISO-C that may have problems if specified in the new Standard.
Reassigned to Ulrich Drepper
A9903-06 AJ (as PASC study group chair) to ensure that all projects that are within the scope (or possible scope; e.g. projects that may fail to reach the deadline) realize that they should be involved in the Austin Group work, and are appropriately invited to attend.
CLOSED
A9903-07 AJ to check SUD to see if there are good examples that could be pulled in to the "Examples" sections.
CLOSED - there are examples that should be included in a later draft (or possibly in XRAT)
A9903-08 AJ to add Peter Anvin to the Austin group reflector
CLOSED
A9903-09 FP to obtain a draft of 1003.1n for editorial information (not draft 1 inclusion) ASAP
CLOSED
A9903-10 AJ to update Lowell on current negotiations with IEEE
CLOSED
A9903-11 AJ to request a special plenary of SSWG on Wednesday evening at Charlotte
CLOSED
A9903-12 AJ to alter scope to correct UN, PI, OF, OP marked sections of XCU, and OH sections of XSH.
CLOSED
A9903-13 AJ to circulate documents from this meeting ASAP
CLOSED
A9903-14 AJ to provide DWC with updated PAR (PARs) in time for the PMC meeting.
CLOSED
A9903-15 AJ to discuss with Lowell meeting fees for the Montreal meeting
CLOSED
2. Procedures Update
Andrew Josey gave a brief overview of the JPC procedures (JDOCS). These have not yet been finalized, but are expected to be soon. He also gave an overview of the committee draft development procedures (AUSTIN/26). We need to improve our use of Aardvark in order to ensure that opinions and comments are not lost; the austin-review mail list is not intended as a general discussion list, although it has been used as such.
3. Identification of ORs and Chair Election
PASC - Don Cragun.
WG15 - Nick Stoughton.
TOG - Jim Zepeda.
4. Status Reports
4.1 Austin Group Status
Andrew Josey gave a brief overview (AUSTIN/28) of the mission and purpose of the Austin Group as an introduction for the new members. Currently 126 participants (up by 50 since the previous meeting), from a wide mix of interest groups.
Draft 1 published June 1 1999.
Joint Memorandum of Understanding (MoU) signed between IEEE and TOG to cover copyright.
A substantial amount of time has been devoted so far to this standard, and it might be interesting to keep a tally of this for later publication. ACTION Chair to set up a time-counter on the web site.
The MoU defines the base documents, the bylaws and procedures governing the production of the new standard. There will be joint ownership of the resulting work to be shared between the IEEE and The Open Group. Publication and Distribution rights are described, but there was no agreement on free web access to the resulting documents. This is still open for further discussion. Andrew Gollan asked if there were plans to push this issue forward. Andrew Josey stated that this would be progressed. Keld asked about ISO's involvement. Lowell stated that this should be done in the same way that other POSIX standards are handled; once completed ISO will get the documents for free to sell in their own way.
Before publication, electronic access to the documents is permitted, but this will be removed at publication.
4.2 PASC/POSIX Status
Revision PAR approved at June Standards Board. This will be called P1003.1 from the IEEE point of view. IEEE Std 1003.2 will be withdrawn upon approval of this standard. Where a working group has liaison by common membership (as is the case here), the liaison members must be suitably accredited.
ACTION ORs to provide letters authorizing them as representatives of their organization.
We need to merge/include a number of in-flight amendments.
.1a: Nick Stoughton: about to go for final recirculation; needs work on namespace reservation. Should finish very soon. Needs FPDAM ballot in ISO.
.1d: Frank Prindle: currently in recirculation. Mainly editorial changes. Expects to go to standards board for final approval soon. It has not had FPDAM ballot in ISO yet.
.1g: Lowell: nothing has changed. Craig Meyer, the ballot coordinator, does not have the funding to complete this until next year. This draft has been languishing for several years. It may well be killed soon if no progress can be made.
.1h: Helmut Roth: SRASS hope to get a draft to the IEEE this week for initial ballot. This document is behind schedule, and is extremely unlikely to make the deadline. ISSUE what to do with this project.
.1j: Frank Prindle: similar to .1d, but one more recirculation needed. Concerns over scope limitations that will probably upset some of the ballot group. Close to the wire, but still expected to be included.
.1m: Helmut Roth: Checkpoint - restart. First ballot got 60% approval. Two more recirculations needed, and FPDAM. Unlikely to make it.
.1q: Frank Prindle: Trace: this in first recirculation, and is at 64% approval. Several small changes underway, new objections are expected. Unlikely to make it.
.2b: Don Cragun/Nick Stoughton: in first ballot. One or two more recircs likely. In danger. FPDAM started.
4.3 IEEE/TOG Status
See MoU details above.
4.4 C9x Status
The FDIS vote has not yet happened, but the draft has been delivered to JTC1 for this to start. The FDIS vote is 2 months, yes/no with no comments. It is likely that this will pass.
4.5 Networking Status
Andrew Gollan (Vice Chair of XNET) reported on current status. There is an open technical mailing list (xnet@opengroup.org ) for anyone to join. Drafts are online. Austin Group participants are actively encouraged to participate. XNS (Unix Networking Services) includes sockets and XTI, and is in ballot (closes July 30) with issue resolution soon after. TOG Company Review expected in September. Final meeting in October. Publication after November meeting of Architecture Board. Ready for inclusion in Austin documents this year. XTI will be optional, and maybe deprecated or obsolete. Basic IPv6 interface added. No specification of either Unix domain or raw sockets. Advanced IPv6 interfaces will not be included (not finished at IETF). The getattrinfo() interface from .1g had IPR issues, which have now been resolved, and this interface will be included.
The Austin Common Standard will supercede XNS when it is finally published. Therefore, XNS will probably never be formally printed. It will be available on the web. Should the XNS document be closed off as described, or held open? Some discussion. ISSUE should XNS should be completed and closed off by the end of this year, or remain uncompleted?
There should not be a product standard based on XNS 5v2. This group does not have the power to decide this issue, and while there is a single strong objector (who does have a mechanism for raising his concerns in TOG) this should not go on the issues list, and is not felt generally to be a problem we need to be concerned with.
Should XTI be marked optional or obsolete? If it is marked as obsolete, it will probably not be included in the Austin standard. There is consensus in XNET to mark it as optional. There is not consensus yet that this should be marked legacy or obsolete. In order to exclude it automatically from XSH6 it should be marked as one of these. Probably legacy.
There is much good verbiage in 1003.1g, whether it completes or not. Having two competing standards (XNS and 1003.1g) is not good. 1003.1g covers raw sockets and Unix domain sockets. What should we do to the scope to include these? ACTION Andrew Gollan/XNET to make a proposal on how to include relevant pieces of P1003.1g in our scope; this proposal should be submitted after the October 1999 XNET meeting.
XNS makes reference to various RFCs. How should we make normative references to such standards? ISSUE how to make normative references to non-ISO standards.
4.6 JPC Status
Roger Martin reported that there would be an electronic discussion and ballot on the JDOCS procedures in the next 4 - 5 weeks.
5. Old Business
None.
6. New Business
Aardvark discussions are captured in the aardvark itself, and not further minuted. Only issues and actions arising from these discussions are noted below.
6.1 XBD Aardvark
138 comments/objections.
OPEN - ERN 3, 6
ISSUE: come up with consistent names for the documents to use throughout, and a name for the set.
ACTION: H P Anvin to prepare aardvark providing mapping between names used in the text and the names used in ISO 10646 (see ERN 26).
OPEN ERN 32
ACTION: J Zepeda address ERN 38. Done - accept as marked, change to "Memory associated with a process than can contain dynamically allocated data".
ISSUE - Home directory - need to be very careful about the use of this term. Check all uses of "Home Directory" to ensure that it is used in the context of this definition (changed to use the "user database"), and not used where $HOME should be used instead. ERN 58
No consensus on ERN 59.
ACTION: Frank Prindle to check on use of "System Scheduling Priority" in XCU to see if the term can be replaced with "Nice Value". (ERN 81)
ACTION: Mark Brown to contact Gary Miller for rationale on positive characters (ERN 93/94) and propose aardvark changes to include rationale for this decision.
ACTION Mark Brown to contact Gary Miller regarding ERN 112 to propose appropriate course of action.
ACTION Nick Stoughton to clarify position on Canadian Std Z234.4.1-1990 (is it proposed, adopted, revised?)
ACTION Mark Brown to contact Gary Miller regarding ERN 120 to propose appropriate course of action. Ulrich Drepper to be included in this discussion.
ACTION Peter Anvin to check proposed action for ERN 130
ISSUE index - what to do with cross volume references, if a common index to all volumes is required. Also address all issues in ERN 136.
6.2 XCU Aardvark
376 aardvark enhancement requests submitted.
ISSUE: what version of C is this standard to be based on. If it is c99, should c89 utility refer to the c99 compiler, or will there be a new utility?
ISSUE: ERN 157 If c89 is kept, but we align with c99 XBS5 is not an appropriate identifier, since it refers to a different version. (see also getconf)
ACTION: TOG OR to investigate the problems with c9x and XSI conformance.
ACTION: Peter Anvin to rewrite the calendar section to handle the Julian/Gregorian switchover problem.
ISSUE ERN 195 cannot be resolved until c9x issues are closed. Since c9x strftime and POSIX locale definition of %c and %p differ, we cannot simply refer to strftime at this point.
ACTION Andrew Josey to file an interpretation request against dd (in POSIX.2) regarding swan inconsistency (ERN199)
ACTION: WG15 OR to request information on alternative dd command syntax (see ERN 295)
ACTION: All: review all ex and vi aardvarks after .2b merge.
ACTION: Donn Terry to investigate historic practice regarding ERN 285 and check that his proposed action is reasonable. Similarly for 286, proposing a suitable action.
ACTION: Nick Stoughton to ask Eric Allman for advice on ERN 288
ISSUE Need to check all uses of "nice value" and "system scheduling priority" in XCU, XSH, and XBD (see action arising from XBD above).
ACTION Andrew Josey to check on ERN 335 resolution.
6.3 XSH Aardvark
448 aardvark enhancement requests submitted.
ISSUE - as above for version of C to be used. If c99, then how to get new functions, how to change footprint of existing functions etc.
ACTION Keld Simonsen to conduct informal discussion with WG14 and ISO central secretariat regarding using c99 material in a guide and copyright issues in including c99 functions in this standard.
ISSUE need to look at all informative text to find normative requirements, and to move them appropriately.
ISSUE need to get a better conformance statement for entire document set
ACTION: TOG Base WG to address ERN40
ACTION Frank Prindle to assist the editors with shading in Realtime section.
ACTION Donn Terry to propose alternative wording for ERN 49
ACTION Donn Terry to file an interpretation request for ERN 51
ISSUE what to do about 1003.1n? Should it be balloted, or should it die and the PAR for this project be modified to include that work?
ACTION TOG Base WG to consider bsd_signal and decide if this should become legacy.
ACTION Peter Anvin to write definition for ISVTXT on directories for XBD (and possible changes for chmod in XCU)
ACTION Frank Prindle to check all EIO errors as defined in ERN 82
ACTION TOG Base WG to consider appropriate action for ERN 90
ACTION TOG Base WG to consider appropriate action for ERN 92
ACTION TOG Base WG to consider appropriate action for ERN 94
ACTION Frank Prindle to check shading on errno for ERN 102
ACTION Peter Anvin to direct a proposal for #! To OGTGBASE for their consideration
ACTION TOG Base WG to consider appropriate action for ERN 111
ACTION Ulrich Drepper to provide example of getsubopt for ERN 157
ACTION TOG Base WG to consider appropriate action for ERN 168 and 169
ACTION TOG Base WG to consider appropriate action for ERN 190 and 191
ACTION TOG Base WG to consider appropriate action for ERN 219 and 225
ACTION TOG Base WG to consider appropriate action for ERN 220
ACTION Donn Terry to submit an interpretation request against POSIX.1 regarding ERN 249
ACTION TOG Base WG to consider appropriate action for ERN 259 and 260
6.4 Style Issues
Tuesday: attempted conf call but with no speakerphone. Donn Terry present. This was abandoned as unworkable after a few minutes.
Nick Stoughton went and bought a speakerphone - thanks to Usenix for this!
Wednesday: Donn Terry joined by conference call (after 31 minutes waiting for him).
1. Use 'shall' not 'will' or present tense. Original POSIX editors found that 'shall' was very imposing, and tried to make the documents more human readable. However, there was substantial pushback from IEEE and ISO. AGREED Use of the imperative should be restored and included in the new text. Go through POSIX.1 and POSIX.2 to find all uses of "shall" (and will), [plus maybe "should", "can", "must", "may", "might" etc] and restore this usage in normative context. Use the same style in XSI and MAN shaded sections.
2. Character quoting: macro-ize all uses of C char constants (single quote CW), C strings, (double quoted CW), shell operators (bold CW, no quotes) etc. ACTION - Nick Stoughton to send Cathy Hughes the relevant part of the POSIX macros and PSSG text
3. Volume naming; ACTION Yvette Ho Sang: Check if we can define abbreviation "XBD" to "The XBD part of 9945-1" etc, and then use this abbreviation everywhere a cross volume reference is required.
4. "Man Page" Style: ACTION Yvette Ho Sang to check on the heading/clause numbering requirements both at IEEE and ISO (we want no numbering, and certainly forced page-breaks).
6.5 Global Issues
Mostly covered in style above.
7. Consent List
The following items were agreed upon during this meeting and added to the Consent list (SD/2)
· AGREED Use of the imperative should be restored and included in the new text. Go through POSIX.1 and POSIX.2 to find all uses of "shall" (and will), [plus maybe "should", "can", "must", "may", "might" etc] and restore this usage in normative context. Use the same style in XSI and MAN shaded sections. (Note that this is for normative text only).
8. Issues List
ISSUE: come up with consistent names for the documents to use throughout, and a name for the set.
ISSUE index - what to do with cross volume references, if a common index to all volumes is required. (see also XBD D1 ERN 136).
ISSUE: what version of C is this standard to be based on. If it is c99, should XCU c89 utility refer to the c99 compiler, or will there be a new utility? If c99, then how to get new XSH functions included, how to change footprint of existing functions etc.
ISSUE: (see XCU D1 ERN 157) If c89 is kept, but we align with c99 XBS5 is not an appropriate identifier, since it refers to a different version. (see also getconf)
ISSUE (see XCU D1 ERN 195) cannot be resolved until c9x issues are closed. Since c9x strftime and POSIX locale definition of %c and %p differ, we cannot simply refer to strftime at this point.
ISSUE need to get a better conformance statement for entire document set
ISSUE what to do about 1003.1n? Should it be balloted, or should it die and the PAR for this project be modified to include that work?
9. Closing
9.1 Review of Action Items
9907-01. Mark Brown to add columns to AUSTIN/6 for what .13, FIPS, & SUS require to be able to control.
OPEN
9907-02. Andrew Josey to provide Keld Simonsen with some examples of how a function (e.g. fseek) might look in time for the next WG14 meeting in London, WB 21 June.
LEAVE OPEN - change to October 1999 meeting
9907-03. Keld Simonsen to supply a list of overlapping functions between XSH and ISO-C that may have problems if specified in the new Standard.
Reassigned to Ulrich Drepper
9907-01. Nick Stoughton to send Cathy Hughes the relevant part of the POSIX macros and PSSG text
9907-02. Yvette Ho Sang: Check if we can define abbreviation "XBD" to "The XBD part of 9945-1" etc, and then use this abbreviation everywhere a cross volume reference is required.
9907-03. Yvette Ho Sang to check on the heading/clause numbering requirements both at IEEE and ISO
9.1.1 Actions Arising from XBD Aardvark
9907-04. H P Anvin to prepare aardvark providing mapping between names used in the text and the names used in ISO 10646 (see ERN 26).
9907-05. J Zepeda address ERN 38. CLOSED - accept as marked, change to "Memory associated with a process than can contain dynamically allocated data".
9907-06. Frank Prindle to check on use of "System Scheduling Priority" in all volumes to see if the term can be replaced with "Nice Value". (ERN 81)
9907-07. Mark Brown to contact Gary Miller for rationale on positive characters (ERN 93/94) and propose aardvark changes to include rationale for this decision.
9907-08. Mark Brown to contact Gary Miller regarding ERN 112 to propose appropriate course of action.
9907-09. WG15 Nick Stoughton to clarify position on Canadian Std Z234.4.1-1990 (is it proposed, adopted, revised?)
9907-010. Mark Brown to contact Gary Miller regarding ERN 120 to propose appropriate course of action. Ulrich Drepper to be included in this discussion.
9907-011. Peter Anvin to check proposed action for ERN 130
9.2 Actions Arising from XCU Aardvark
9907-012. TOG OR to investigate the problems with c9x and XSI conformance.
9907-013. Peter Anvin to rewrite the calendar section to handle the Julian/Gregorian switchover problem.
9907-014. Andrew Josey to file an interpretation request against dd (in POSIX.2) regarding swab inconsistency (ERN199)
9907-015. WG15 OR to request information on alternative dd command syntax (see ERN 295)
9907-016. Donn Terry to investigate historic practice regarding ERN 285 and check that his proposed action is reasonable. Similarly for ERN 286, proposing a suitable action.
9907-017. Nick Stoughton to ask Eric Allman for advice on ERN 288
9907-018. Andrew Josey to check on ERN 335 resolution.
9.3 Actions Arising from XSH Aardvark
9907-019. Keld Simonsen to conduct informal discussion with WG14 and ISO central secretariat regarding using c99 material in a guide and copyright issues in including c99 functions in this standard.
9907-020. TOG OR to pass to TOG Base WG to address ERNs 40, 49, 51, 82, 90, 92, 94, 111, 168, 169, 190, 191, 219, 225, 220, 259, 260
9907-021. Frank Prindle to assist the editors with shading in Realtime section.
9907-022. Donn Terry to propose alternative wording for ERN 49
9907-023. Donn Terry to file an interpretation request for ERN 51
9907-024. TOG OR to pass to TOG Base WG to consider bsd_signal and decide if this should become legacy.
9907-025. Peter Anvin to write definition for ISVTXT on directories for XBD (and possible changes for chmod in XCU)
9907-026. Frank Prindle to check all EIO errors as defined in ERN 82
9907-027. Frank Prindle to check shading on errno for ERN 102
9907-028. Peter Anvin to direct a proposal for #! To OGTGBASE for their consideration
9907-029. Ulrich Drepper to provide example of getsubopt for ERN 157. CLOSED
9907-030. Donn Terry to submit an interpretation request against POSIX.1 regarding ERN 249
9.4 Document Register
Note that Austin/SD1 (Issues list) and Austin/SD2 (Consent list) will be updated based on these minutes.
AUSTIN/28 |
Andrew Josey |
Status Presentation |
AUSTIN/29 |
Nick Stoughton |
Minutes of the July 1999 Meeting |
9.5 Next Meeting(s) / Future plans
The remaining aardvarks will be checked for clarity and precision, and if they are good enough we will vote on them electronically.
A sanity check draft (2) will be made available as soon as the aardvark analysis has completed and the required edits made. This is expected by the end of September/beginning of October.
Draft 3 will include XNS 5 release 2, 1003.1a, and 1003.2b, plus those amendments ready for editors to include by 12/1/99. Draft 3 should be available by 2/29/00. There will be an 8 week review period for the draft 3 review.
New timeline:
October 1999 Draft 2 available (no formal review period)
Feb 29 2000 Draft 3 available
Mar 1 2000 - April 30 (8 weeks) Draft 3 review
May 15-21 review meeting
The next meeting of the Austin Group is likely to be a planning meeting in late December 1999 or early January 2000 to decide contents for draft 3 and discuss any issues such as c9x.
Next review meeting in May 2000 in Cupertino or Copenhagen. (Date/place to be announced).