From baker@ada.cs.fsu.edu Wed Jul 10 20:31:28 1996 Received: from ada.cs.fsu.edu (ada.cs.fsu.edu [128.186.121.48]) by dkuug.dk (8.6.12/8.6.12) with ESMTP id UAA20916 for ; Wed, 10 Jul 1996 20:31:24 +0200 Received: by ada.cs.fsu.edu (SMI-8.6/SMI-SVR4) id SAA25879; Wed, 10 Jul 1996 18:30:36 GMT Date: Wed, 10 Jul 1996 18:30:36 GMT Message-Id: <199607101830.SAA25879@ada.cs.fsu.edu> From: Ted Baker To: SC22WG15@dkuug.dk, wg15tag@pasc.org Subject: liason from WG9 The following item came over the WG9 e-mail list, and may be worth the attention of WG15 also. --Ted Baker | From: demery@CCGATE.HAC.COM | Date: Wed, 10 Jul 96 10:40:38 PST | Encoding: 52 Text | Message-Id: <9606108370.AA837020669@CCGATE.HAC.COM> | To: iso@sw-eng.falls-church.va.us | Subject: Concerns on PCTE Ada Binding ballots | Content-Type: text | Content-Length: 2819 | Status: R DISCLAIMER: This is a personal "technical contribution" and not the position of any WG9 member body... I'm quite concerned about the continuing SC22 ballots on Ada Bindings to PCTE. As I recall, the original PCTE/Ada binding was processed via Fast-track procedures through JTC1, as one component of a set of existing PCTE standards. The original PCTE Ada binding was done a very long time ago, and was not consistent with POSIX/Ada bindings covering the same basic service area (particularly signal delivery and handling.) (When we did POSIX/Ada, we looked at the PCTE/Ada approach, and decided that it would not work in all cases. It was one of several alternative signal models that we considered before arriving on the model in POSIX.5 and the revised model in POSIX.5b.) At that time, I expressed my concerns about this, and recommended a No vote on this part of the original PCTE work. My belief was that the PCTE/Ada binding needed to be revised to make it consistent with changes to Ada and Ada bindings 'technology'. Additionally, the PCTE/Ada binding proposal had not undergone very extensive review within the Ada community (unlike the POSIX/Ada binding.) Well, JTC1 voted to approve the initial PCTE/Ada binding, and assigned the work to SC22, which then formed a new WG (WG22) to handle PCTE standards. SC22/WG22 has since brought forward new PCTE standards, including new Ada bindings. I have NOT seen the new PCTE Ada bindings, and I'm very concerned that there seems to be little coordination between SC22/WG9 and SC22/WG22 (and SC22/WG15) for Ada bindings to PCTE. The PCTE work overlaps both the WG9 community of interest, and ongoing work within WG15 for Ada bindings to additional POSIX services. Technically, the issue is interoperability and consistency between various Ada bindings that can be used together, such as PCTE and POSIX. This is less of a problem for the C community, since both POSIX and PCTE "share" a common Unix heritage, and there is widespread existing practice that makes sure that PCTE and POSIX do similar things using similar interfaces. So my personal opinion is that no Ada binding should be approved by SC22 without prior coordination/advise from WG9. Formally, I think that WG9 member bodies should advise their SC22 member body delegates to disapprove PCTE Ada bindings (or other Ada bindings) that have not been reviewed within WG9 and the WG9 member bodies. Frankly, I would be Very Concerned about this if I thought that PCTE was "important". Since I don't see a major move towards PCTE acceptance or use in industry, I personally regard the PCTE standards as mostly academic. But, even for such standards, it is still important that the Ada community ensure that standard Ada bindings are done well. dave