Thread View: comp.std.unix
4 messages
4 total messages
Started by Stephen R. Walli
Tue, 29 Dec 1992 22:26
Standards Update, POSIX.7b: Software Administration
Author: Stephen R. Walli
Date: Tue, 29 Dec 1992 22:26
Date: Tue, 29 Dec 1992 22:26
79 lines
4645 bytes
4645 bytes
Submitted-by: stephe@usenix.org (Stephen R. Walli) Esti Koen <emk@cray.com> reports on the October 19-23, 1992 meeting in Utrecht, NL: I attended the POSIX.7b meeting in Utrecht, never having been previously exposed to POSIX. Lacking the historical perspective, it was difficult for me to identify when the discussion was a clarification of an already agreed upon point verses a major shift in emphasis or direction. If this report seems somewhat lacking in detail or introductory, it reflects my own level of involvement to date. For the purpose of this report, I assume readers are mainly interested in broad decisions concerning the content of the standard or a shift in direction and expected balloting dates. Early attempts to standardize the nonexistant ``common practice'' of software administration seemed doomed to failure. (I don't envy those early pioneers.) POSIX.7 finally adopted the network view of a managed system. Forging ahead in areas where they feel they can make consensus based progress, POSIX.7 is now split into two documents called POSIX.7a (print queue administration) and POSIX.7b (software administration). Recognizing the need for information describing existing practice in the area of network wide system management, the Open Software Foundation (OSF) solicited technologies from industry that could be integrated to simplify system management in heterogeneous computing environments. In October, 1991, OSF announced that they had chosen Hewlett Packard's Software Distribution Utilities to provide the basis for the OSF Distributed Management Environment (DME). The current draft of POSIX.7b is a roughly one year old descendant of the External Specification that describes the HP Software Distribution Utilities. The original HP implementation suggested an object orientation but it was not developed using a rigorous object oriented specification language. In one year of POSIX meetings the group has made significant progress in further defining the attributes of the managed objects, but the specification is still incomplete and at times ambiguous. There is much discussion concerning object behavior. Open issues include the question of allowing multiple Management Information Bases (MIB), and which attributes of a software object can be used, and how they are used as a selection mechanism. Although invention by a standards committee is not advisable, it seems unavoidable when the base design is incomplete for the purposes of the standard. Several decisions regarding general content were finalized. There will be no API included in the standard. An informative annex which provides information on how one implementation communicates between the manager, source and target roles will be included. A rationale section which informs the reader as to the intent and history of the standard will also be included. The serial media format was previously specified as tar, but will now be specified as being readable and writable by pax (POSIX.2-1992). Locking mechanisms are considered to be an implementation detail and outside the scope of the standard. A command line option will be provided to permit interaction sufficient to handle multi-volume media. The group discussed rewriting part of the document using the ISO Guidelines for the Definition of Managed Objects (GDMO). The process of rewriting using GDMO would have the beneficial side effect of highlighting inconsistencies, omissions, and redundancies. In fact, it was advised that the draft would not be adopted by ISO unless GDMO was used. The active participants did not embrace the idea whole heartedly because a drastic structure change could further delay the balloting schedule. Mock ballot is planned to occur after the January meeting. Budget constraints may impose a time limit on the standards activity, and active participants fear having the POSIX.7b standards activity permanently interrupted before going to ballot. Refinement of the existing object definitions and behaviors continues at a fast pace. Volume-Number: Volume 30, Number 12
Re: Standards Update, POSIX.7b: Software Administration
Author: peter da silva
Date: Tue, 05 Jan 1993 00:27
Date: Tue, 05 Jan 1993 00:27
26 lines
1628 bytes
1628 bytes
Submitted-by: peter@ferranti.com (peter da silva)
In article <1hqj9sINN9mu@ftp.UU.NET> std-unix@uunet.uu.net writes:
> Forging ahead in areas where they feel they can make
> consensus based progress, POSIX.7 is now split into two documents
> called POSIX.7a (print queue administration) and POSIX.7b (software
^^^^^^^^^^^^^^^^^^^^^^^^^^
> administration).
Isn't this a rather narrow subject for a POSIX standard... not to mention
being pretty much covered by two completely irreconcilable models (BSD
and System V).
You're going to have to dump BSD spooling, dump System V spooling, or
dump both and create something out of whole cloth. If it's the latter,
why not create a generalised batch queueing and scheduling system that
just happens to include printing?
I assume 7b is where all the interesting stuff is going to be.
--
Peter da Silva `-_-'
Ferranti International Controls Corporation 'U`
Sugar Land, TX 77487-5012 USA
+1 713 274 5180 "Zure otsoa besarkatu al duzu gaur?"
Volume-Number: Volume 30, Number 16
Re: Standards Update, POSIX.7b: Software Administration
Author: Duke Robillard
Date: Sun, 10 Jan 1993 11:46
Date: Sun, 10 Jan 1993 11:46
37 lines
2070 bytes
2070 bytes
Submitted-by: duke@portal.paperboy.osf.org (Duke Robillard) In article <1iakluINNblv@ftp.UU.NET> peter@ferranti.com (peter da silva) writes: >>consensus based progress, POSIX.7 is now split into two documents >>called POSIX.7a (print queue administration) and POSIX.7b (software > ^^^^^^^^^^^^^^^^^^^^^^^^^^ >administration). >Isn't this a rather narrow subject for a POSIX standard... >I assume 7b is where all the interesting stuff is going to be. Actually, POSIX.7 is being split into n documents, one for each area of system administration. 1003.7.1 (what Esti called .7a) is about print systems, 1003.7.2 (formally known as .7b) is about software package installation, 1003.7.3 will be about user management, etc. >You're going to have to dump BSD spooling, dump System V spooling, or >dump both and create something out of whole cloth. Well, some people claim we've done the later, although that a little extreme. We've adopted Palladium, the print system developed at MIT's Project Athena. It's an implementation of the ISO Print Standard. >If it's the latter, why not create a generalised batch queueing and >scheduling system that just happens to include printing? We thought about that, but couldn't find any software like that to base the standard on. -- Bob Robillard Technical Editor, 1003.7.1 duke@osf.org Duke Robillard, DoD #0130, duke@osf.org angst.motorcycles: "We may be doomed, but at least the adrenaline keeps us awake" Volume-Number: Volume 30, Number 19
Re: Standards Update, POSIX.7b: Software Administration
Author: Ran Atkinson
Date: Sun, 10 Jan 1993 12:32
Date: Sun, 10 Jan 1993 12:32
20 lines
1378 bytes
1378 bytes
Submitted-by: atkinson@tengwar.itd.nrl.navy.mil (Ran Atkinson) Doug Gwyn <gwyn@smoke.brl.mil> has been using such generalised batch software on UNIX systems. I think his stuff might even be freely distributable. I find it absolutely fascinating that you continue to claim that MIT developed Palladium and yet the vast majority of MIT won't have anything to do with it -- including large portions of Project Athena. (We sent someone from NRL up to MIT to survey the Palladium situation :-) To me that is a clear sign that there are problems with Palladium. Also, existing practice at one site out of hundreds of thousands is not anything resembling generalised existing practice (which both BSD and System V do strongly resemble). Ran atkinson@itd.nrl.navy.mil Volume-Number: Volume 30, Number 23
Thread Navigation
This is a paginated view of messages in the thread with full content displayed inline.
Messages are displayed in chronological order, with the original post highlighted in green.
Use pagination controls to navigate through all messages in large threads.
Back to All Threads