jsh@usenix.org (Jeffrey S. Haemer) (08/16/90)
From: Jeffrey S. Haemer <jsh@usenix.org> An Update on UNIX*-Related Standards Activities August, 1990 USENIX Standards Watchdog Committee Jeffrey S. Haemer <jsh@usenix.org>, Report Editor USENIX Standards BOF An anonymous correspondent reports on the June 12 meeting in Anaheim, California: If they find out who I am... The snitch requests anonymity for several reasons, none of them related to his alcohol consumption during the bof. (No officer, I swear I wasn't going to log in and do system administration until I sobered up.) The request actually relates to the snitch's employer -- a standards organization. Because I am paid neither to file snitch reports nor to write opinions on standards, to submit this paper through normal channels for official, outside publication, even if it were entirely objective (or factual, for that matter), would require endless rounds of exhaustive, organizational review. On to the meeting. As usual, the meeting was held immediately after the official USENIX reception, which meant that the snitch continued to suck down his third or fifth beer as the meeting opened. John ``standards is politics'' Quarterman, of Texas Internet Consulting (TIC), and Susanne Smith, of Windsound, chaired the meeting, which was attended by about 40 people, including Larry Wall -- nearly a standards body by himself. [ Editor: Larry is the person responsible for such contributions to the community as rn, patch, and perl. ] Jeff Haemer was absent because ``his wife is having a baby any day and I just don't know where his priorities are!?'' [Editor: Zoe Elizabeth Haemer, 6lbs. 10oz., after a forty-five minute labor] John started out by covering the usual stuff -- who he is, how to reach him, what he does, [Editor: Sounds like it would have been valuable for me to attend.] and so on. You should already know all this since it is covered regularly in articles in the publication or newsgroup in which you reading this article. John gave some updates __________ * UNIXTM is a Registered Trademark of UNIX System Laboratories in the United States and other countries. August, 1990 Standards Update USENIX Standards BOF - 2 - for things that are probably already out-of-date, so I won't repeat them. Susanne pointed out that TIC and Windsound have collaborated on a calendar that includes all the latest dates of standards meetings, which they were giving away for free at the meeting. [Editor: You can request copies from tic@tic.com. They span July 1990-June 1991, and cost $5.00, plus shipping, handling, and (Texans only) tax.] John and Susanne briefly reviewed standards efforts of interest to USENIX members, including P1003 (POSIX) and P1201 (Windowing). John discussed whose standard (ISO? ANSI? FIPS? other?) was most important but I was unable to draw any conclusions or coherently summarize it, so I'll omit it here. Nonetheless he did get across two points: 1) there is a lot of coordination between groups and 2) he is very quotable. (``The IEEE standards board is baroque and byzantine.'') The crowd becomes surly After this basic informational introduction, the meeting was thrown open to the audience. The ensuing discussion was a mix of four things: 1. Humor A couple of examples will give the flavor. + An overheard conversation: ``Mach was the greatest intellectual fraud in the last ten years.'' ``What about X?'' ``I said intellectual.'' + The announcement of the new Weirdnix contest: a contest for a correct interpretation of P1003.1 or .2 furthest from the original intent. The state of Utah (I am not making this up) is offering a trip for two to Salt Lake City for the winner. 2. Opinion polling John tried to discern whether attendees thought they were being well-served by John, the USENIX Standards Watchdog Committee, and the USENIX position on standards: to attempt to prevent standards from prohibiting innovation. Indeed, at Snowbird, the site of the April POSIX meeting, John was told that smaller companies don't like our participation because of this position. Think about this a while. (For a more detailed discussion of the USENIX position on standards, see either ;login: 15(3):25 or August, 1990 Standards Update USENIX Standards BOF - 3 - the periodic overview posting in comp.std.unix about the USENIX Standards Watchdog Committee.) John explained how USENIX came to its current policies and why it does not endorse standards of its own. Some audience members were unhappy with extant standards bodies and said they wouldn't mind if USENIX played a more active role. Susanne reminded us that UniForum working groups, which she praised, play such a role. You are encouraged to tell John and the USENIX Board what you feel the USENIX position on standards should be, how much money USENIX should budget for standards activities, or anything else that's on your mind. (The current USENIX standards budget is $45K/yr.) On a related note, BOF attendees were quite eager to be kept informed on standards issues. In the snitch's opinion, this is probably the standards-related area in which USENIX most excels, and its contribution overshadows that of any other source that this snitch is aware of. The USENIX Standards Watchdog Committee publishes copiously in both ;login: and the usenet newsgroup comp.std.unix. (The level of detail can certainly not be said to be too high, but USENIX Board meetings continually propose reducing it.) While the newsgroups get the information more quickly, ;login:, in particular, remains the official voice of USENIX, and standards issues now fill 1/3 to 1/2 of each edition. Many non-UNIX aficionados who want to stay current on related standards join USENIX simply to get ;login:. Both John and the Board believe that although the newsgroup has been quite active this past year, hard copy still circulates more widely. Some attendees wanted increased coverage of standards currently outside of ;login:'s bailiwick, such as RS-232 and CD-ROM format. Unfortunately, following any and all computer-related standards would exceed USENIX's budget and resources. [Editor: The alert reader will have noticed Andrew Hume's fine report on WORM-based file system standards last quarter. Send me a report. I'll edit it. ] John raised the possibility of breaking out the standards information of ;login: into a separate publication. This was also discussed at the USENIX Board meeting during the week. Stay tuned. John and Susanne revealed that they are writing a book on UNIX- related standards (which will not be posted electronically). No suggestion was made for how it could possibly stay up to date. August, 1990 Standards Update USENIX Standards BOF - 4 - 3. Government-bashing (Who the hell is NIST and why are they so out of control?) As soon as we determined that NIST wasn't represented in the room and couldn't defend itself, it became fair game. (There were no OSF reps either -- their BOF ran concurrently with ours -- but no one knew what OSF was doing so we skipped insulting them.) John fanned the flames by giving an example where NIST had pushed too hard, in his opinion: System Administration. ``Dot seven shouldn't exist,'' he said, but NIST pushed for it. Because government agencies view FIPS so favorably that a system administration FIPS would quickly become a de facto standard for non-government users as well, the IEEE said ``ok, let's look at it.'' John said things didn't turn out as badly as they could have. Unfortunately there is little common practice or prior art in the area; fortunately, dot seven is coming along so slowly that there may be by the time it is ready to go to ballot. Moreover, dot seven's work has encouraged several companies and universities to work on the parallels between system administration and network management. Still, he reminded us that a standard should neither create nor innovate but only standardize, quoting Dennis Ritchie's compliment to X3J11 in his keynote address: ``The C committee took something that wasn't broken, and tidied it up without breaking it.'' The audience asked, ``How do we control the activities of NIST?'' NIST is a part of the government. If you are a U.S. citizen, your tax dollars fund it, so you can write your congressperson. While you can communicate directly with NIST's standards representatives, John asked that we not bug them in the name of USENIX, ``because I have to work with these guys.'' If you feel bold, you can actually talk to John Lyons, the director of NIST -- <lyons@micf.nist.gov> -- who lies midway between the scutpuppy standards reps and the demonically powerful congresscritters. He really does read and answer his email (and his signature does say that his opinions represent those of his organization). John ended by defending, or at least rationalizing, NIST's pro- active stance: ``The primary reason is money.'' A familiar example is the Air Force's AFCAC-251 RFP (Request For Purchase). This five-to-ten-billion-dollar request for SVR3-conforming systems created a heap of trouble by specifying a vendor brand name. After official protests, the procurement had to be reworded at great expense -- ultimately to you, the taxpayer. A vendor-independent, POSIX FIPS would have prevented this. August, 1990 Standards Update USENIX Standards BOF - 5 - One of the few questions John couldn't answer was, ``Why did NBS change its name anyway?'' This snitch scraped away at the dirt and uncovered the explanation: The U.S. Department of Commerce under which NBS resides had wanted to change the name for many years because NBS has long performed activities quite unrelated to standards. As usual, it was politically bobbled for quite some time until a sufficiently obvious expansion of responsibilities came up for funding at which time (1/89, Reagan) the following announcement was issued: the new name, ``National Institute of Standards and Technology,'' reflects the broadened role and new responsibilities assigned to the agency which will include the traditional functions of providing the measurements, calibrations, data, and quality assurance support to U.S. commerce and industry, together with several new programs to support the aggressive use of new technologies in American industry. NIST's new purpose is ``to assist industry in the development of technology and procedures needed to improve quality, to modernize manufacturing processes, to ensure product reliability, manufacturability, functionality, and cost-effectiveness, and to facilitate the more rapid commercialization ... of products based on new scientific discoveries.'' Several new programs have been created aimed at rapid transfer of technology to U.S. industry. They are: 1. Regional Centers for the Transfer of Manufacturing Technology; 2. assistance to state technology programs; 3. the Advanced Technology Program; and 4. the Clearinghouse for State Technology Programs. Call (301) 975-3058 (NIST Technical Information) if you would like more information on any of these programs or on NIST itself. 4. John's usual exhortation/guilt-trip: get involved in standards! This discussion went on for some time. UNIX is no longer guided by a few bright individuals; it is now in the hands of vested commercial interests, some of which don't give a damn about innovation or good design. August, 1990 Standards Update USENIX Standards BOF - 6 - For the most part, the committees themselves contain intelligent, well-meaning people who really want to create useful standards. But in a small committee, overlooked unintentional flaws can ruin otherwise good work. Snitches help forestall this by functioning as a community ear. If you don't have time to be on a committee, get on the mailing list and continue to read the newsgroups so you can comment on critical issues when they arise. If you don't, you have have only yourself to blame if the standards come out all wrong. August, 1990 Standards Update USENIX Standards BOF Volume-Number: Volume 21, Number 36
sws@calvin.uucp (Susanne Smith) (08/22/90)
From: sws@calvin.uucp (Susanne Smith) >2. Opinion polling > > John tried to discern whether attendees thought they were being > well-served by John, the USENIX Standards Watchdog Committee, > and the USENIX position on standards: to attempt to prevent > standards from prohibiting innovation. Indeed, at Snowbird, the > site of the April POSIX meeting, John was told that smaller > companies don't like our participation because of this position. > Think about this a while. (For a more detailed discussion of > the USENIX position on standards, see either ;login: 15(3):25 or > the periodic overview posting in comp.std.unix about the USENIX > Standards Watchdog Committee.) > > John explained how USENIX came to its current policies and why > it does not endorse standards of its own. Some audience members > were unhappy with extant standards bodies and said they wouldn't > mind if USENIX played a more active role. Susanne reminded us > that UniForum working groups, which she praised, play such a > role. What I remember about this portion of the bof is a little bit different. John had been talking about the new procedures developed by the TCOS (IEEE Computer Society Technical Committee on Operating Systems) SEC (Standards Executive Committee) to limit the proliferation of standards groups and to make sure that groups in existence made acceptable progress (see Volume 19, Number 97). The question was asked as to what other forums are available for groups that do not gain SEC approval. The UniForum Technical Committees were mentioned as one forum for work that might be premature or inappropriate for TCOS. Volume-Number: Volume 21, Number 55