Check out the new USENIX Web site.
StandardsUSENIX

 

A Message from the Chair

Lowell Johnson <lgj@rsvl.unisys.com>, Chair of PASC, reports at the July 1997 meeting in Nashua, NH.

Changes Coming in the Standards Arena

There are several important changes coming in the near future, both to POSIX and IEEE standards in general. POSIX is the more important to this audience; I'll discuss changes to it before the broader changes.

The July meeting of the Portable Application Standards Committee (PASC) Sponsor Executive Committee (SEC) was specifically extended for a special discussion about the future of POSIX standards. This discussion was instigated by a request from HP that, simply stated, requested that PASC no longer be allowed to generate additional POSIX standards. Although ludicrous on the surface, underneath there was some valuable content.

The initial reaction was that this was a cheap shot by HP to save money by reducing the number and complexity of the standards they thought their customers would require them to support. This appeared to irritate a lot of people, especially because HP had previously admitted that POSIX had generated a very large dollar volume of sales for them. This was exacerbated when the HP person making this presentation made sure he was considered an invited guest because HP did not want to appear to be supporting PASC by paying the meeting fees.

However, there was some valuable content underneath the rhetoric and wounded feelings. The name POSIX has considerable value in the marketplace. It was HP's contention that this value was in the core standards, but that most of the current work was aimed at extending the standards into specific niches (especially realtime and Ada), which are eroding the name recognition value of the core standards.

After a couple hours of debate, there was some acknowledgment of this position, so some ad hoc groups were formed to make suggestions on how we should proceed. I think some good progress was made at this meeting, but it was tainted by the fact that HP refused to participate in working toward a solution. Some reference was made to lobbing in a grenade, then running.

There are many possible outcomes before this issue is resolved, but I think one of the most likely will be some changes in terminology. The core standards will be given one name (like "Core POSIX" or "POSIX Classic"), and most of the new work will be given a different name. The biggest problem will be how to deal with the upcoming revisions of POSIX.1 and POSIX.2, which will undoubtedly be discussed in great detail at the October PASC meeting in Reno, NV.

Another issue brought up was the standards work in areas that were considered by some to be outdated. Security was the most often used example. This is a much harder problem to solve in general because of the process rules under which we live. If people are willing to work on a standard, and the balloting process approves it, we have no choice but to process it into an approved standard.

The real concern here was that the wrong technology would be standardized: the old instead of the new. There is a simple solution to this particular problem, but it requires the active participation of industry. If people come forward wanting to standardize the new technology and are willing to put in the time and energy to produce it, it will become a standard. If not, it will not. It's that simple.

The more general changes in standards involve high-level policy and procedure changes being planned or made by the IEEE. At the July PASC meeting the IEEE Standards Board vice chair made the first public presentation of the new Standards Association that the IEEE is forming, which will begin operation in January 1998. There are many changes involved with the creation of the SA, but the most significant is that only SA members will be allowed to cast ballots on proposed IEEE standards.

The change is fairly minimal for individual members: the membership fee is only $10 per year if you are a member of IEEE or the Computer Society. Remember, you already have to be a member of one of these to vote on standards now.

The big change is that corporations or other organizations (such as government agencies, educational institutions, or consortia) may also become SA members (at a much higher cost, of course). The plan is that balloting on a draft standard may be done by either the current method of individual voting or by a new method of corporate voting. The method would be selected when the initial PAR (Project Authorization Request) is submitted. If the corporate model is selected, then each organizational entity would get only one vote, but it would be guaranteed that vote, even through the substitution of the original person assigned to ballot.

This is a major change that will take a while to sink in and to get all the wrinkles ironed out. If you have comments or concerns about these changes, please feel free to send me email at <l.g.johnson@ieee.org> [or to me, <nick@usenix.org> ­ ed].

The Standards Board is also making some small changes in the current process that should reduce the effort (and stress) in bringing a standard through the balloting process. At its last meeting the board agreed in principle to allow people to be taken off a ballot group if they request it. This will solve several current problems caused by people not returning their ballots, the most common reason being they changed jobs, are no longer interested, or in some cases, actually died during the balloting process. This is a significant change because the composition of a balloting group has always been immutably fixed.

The standards department staff and several of the individual sponsors (including PASC) have defined fast track processes to speed the standards process for specifications that have already been developed in another forum. Unfortunately, we have not yet had a project to test these defined processes. PASC tried to get Sun to submit some of the Java specifications for fast track in IEEE, but they have chosen to try to go directly into ISO JTC1 with their specifications, so we are still looking for volunteers. If anyone knows of a project that may be suited to this process, please contact me.

 

?Need help? Use our Contacts page.
Last changed: Oct 23, 1997 efc
Standards Index
;login: Index
USENIX home