[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Finishing Process and why you are not done.



Title: 754R Folks,

754R Folks,

 

This is to review the process from where we are now to the IEEE Floating Point STANDARD being published.

 

Now we go to Sponsor ballot, followed by submission to the IEEE Standards Board through RevCom.  Here are the steps:

 

Your action at the June meeting will request that the Sponsor (The MSC and the Chair acting for the MSC) to form a Sponsor Balloting Committee for the p754r project to review and approve the draft document for publication.  This is the formal peer review of the standard that backs up the Standards Board, and the ANSI decision to publish the standard.

 

Following the MSC meeting, I will, or will authorize someone, the more usual case it is the chair of the working group, to initiate the MyBallot action to create the balloting pool with some of the preset balloting pools listed, starting with the MSC.  This initiation will send an invitation to all those in the pool and any additional list to join the ballot. This is an open process and anybody that is an IEEE member, or pays the balloting fee, can participate in the ballot through the MyBallot process on the http://standards.ieee.org/stdsdev/index.html Balloting Center.  This ballot is open to all whom wish to participate.

 

Please note well – a large balloting pool is very difficult to get the required return rate. The ballot fails if there in insufficient returned votes! It is not uncommon for this chasing in of the ballots to be one of the more difficult tasks in the ballot. A lot of folks want the free draft only. My next biggest job will be to get the ANSI required balance on the balloting body.  My request would be that you specifically invite those willing to take the time and trouble to vote and comment.

 

From the IEEE Ops manual:

“Balloters are required to classify their relationship to the balloting group relative to the scope of standards activity (for example, producer, user, and general interest). Where appropriate, additional classifications, such as "testing laboratory" or "academic," may be added by the Sponsor. This decision should be based on the effect the standard may have on participants not already recognized by the primary classifications. Individuals classify themselves based on their technical background, which may be related to their employment, job functions, or experience. IEEE-SA entity balloters are classified based on their entity interest as it relates to the scope of the standards project (for example, producer, consumer, general interest). No group (classification) is permitted to constitute 50% or more of the balloting group membership. Care shall be taken to ensure that all classes of interest are represented to the extent possible.

 

It is desirable to have representation of the materially interested and affected parties when reviewing the balance of the balloting group. Sponsors shall ensure balance prior to conducting a Sponsor ballot. Balloting groups should have at least 10 members to ensure adequate balance.”

 

Section 5.4.1 – “The balloting group shall meet the criteria in subclause 5.2.2.3 of the IEEE-SA Standards Board Bylaws. Balloting group members vote on the acceptability of a new or revised standard based on their interest in and commitment to reviewing and voting on a specific proposed standard. The balloting group shall provide for the development of consensus by all interests significantly affected by the scope of the standard. This is achieved through a balance of such interests in the balloting group membership. Balance is defined as the avoidance of dominance by any single interest category.  The name, affiliation if any, such as employer or sponsor, and interest category of each member of the consensus body shall be made available to interested parties upon request.

 

Unless it is claimed in writing (including electronic communications) by a directly and materially affected party that a single interest category, individual or organization dominated the standards development process, no test for dominance is required.”

 

 

All Negative ballots must be accompanied with the written changes that, if incorporated in the draft, would cause the negative voter to change there position to affirmative!  Any Negative vote without the list of required changes is not a valid vote and will not be counted.

 

There are many reasons to vote negative and it is not uncommon for the Committee chair, or member, to vote negative to incorporate fixes to the draft since is was first offered for ballot!

 

If any negative ballots are returned, they must be re-circulated to all voters to give them the opportunity to agree with the negative ballot and change their vote to negative. This recirculation is after the Ballot Review Committee (BRC) meets to discuss the negative vote and its merits and my change the draft to accommodate the negative vote, hence reversing the vote, or partially accept or reject the negative comments and then re-circulate the modified draft, along with the negative comments and the actions taken on each negative, and positive, comment and the decision of the BRC on each comment.

 

This process can go through many generations. However any voter may not object to any point that was agreed to previously unless it is affected by the negative votes.  This prevents the same comment coming in on each circulation of a ballot.

 

Who is the BRC?  The Sponsor appoints the BRC for each balloting body. The sponsor will also look to the sitting expertise for this group! You are not done yet! The BRC is usually a small group from the working group, a subcommittee, or a committee of the whole working group.  Smaller is better!  This BRC group must formally answer each comment, by accepting it into the draft, rejecting it with reason, or possibly negotiating with the commenter for a mutually agreeable solution. All written correspondence is captured and accompanies the document through the SASB/RevCom process. It will be examined for due process!

 

My best guess is that this will take several to 6 months to complete with multiple re-circulations needed to achieve closure.  Re-circulation ballots are 10 day ballots.  The BRC will probably take more time than this.

 

 

This ballot MUST have a 75% return rate and MUST have 75% passage rate.  Reported like this for a recent ballot.

 

Ballot Open Date: xx/xx/xxxx

Ballot Close Date: xx/xx/xxxx

 

RESPONSE RATE

This ballot has met the 75% returned ballot requirement.

75 eligible people in this ballot group.

64 affirmative votes

2 negative votes with comments

0 negative votes without comments

4 abstention votes

70 votes received

=

93% returned

6% abstention

 

APPROVAL RATE

The 75% affirmation requirement is being met.

64 affirmative votes

2 negative votes with comments

66 votes = 97% affirmative

 

This particular ballot pool was smaller than the one anticipated for 754r.

 

Once the response and approval rates are in the proper range and the negative votes have been answered with the necessary modification of the draft and the recirculation ballots passed, the standard and all balloting related material is forwarded to RevCom for process review including matching the PAR statement with the submitted standard – did we do what we said we were going to do? Please see:

http://standards.ieee.org/board/nes/projects/754.pdf

 

Another remaining issue is to collect all the Copyright License forms.  Start checking the attendance list for the list of needed licenses.

 

Your have all made very good progress, it is not done yet, there are still opportunities to fix things, and more work ahead.

 

Then we can start again……..

 

Respectfully,
 
Bob Davis

Member IEEE p754r working group

Sponsor of p754.
Chairman Microprocessor Standards Committee
Member AudCom, Past Member IEEE SA Standard Board
Summit Computer Systems, Inc.
22685 Summit Road, Los Gatos, CA 95033-9310
bob@xxxxxxxx
408.353.2706 Desk
408.857.1273 Cell
408.353.8116 Fax

 


754 | revision | FAQ | references | list archive