ASQ - Software Division

US SC7 TAG Meeting Report

Scott Duncan, SW Division Standards Chair

This column is being written based on the meeting at Boeing in Long Beach, CA ( March 21-23, 2006) through the efforts of Division member and past chair, Mike Kress.

At this 54 th meeting of the US SC7 TAG, Mike Gayle of JPL, was able to return as Chair. (Mike is also a Division member.).

Task Group Presentations

What follows are excerpts from the “outbriefs” (i.e., final day reports) from each Task Group (TG) which explain the current status of various SC7 standards at the international Working Group (WG) level. TGs exist at the TAG level to track and work on the standards and proposals assigned to the WGs at the international level. In some cases, TAG members and other U.S. Technical Experts participate as WG Chairs or Editors

Unless otherwise noted, the commentary relates to the international status of the standards discussed. When US TAG-specific activities are reported they will be so noted.

System and Software Documentation (TG 2)

ISO 15289 ha been approved at the FDIS (Final Draft International Standard) level which means it will soon be published as an IS (i.e., a full International Standard).

Software Product Quality Measurement (TG 6)

There are six documents from this WG up for one form or another of ballot over the coming 4-6 weeks before the annual Plenary meeting of SC7 in Bangkok. The U.S. continues to vote to disapprove all the documents, not because it objects to a set of product quality characteristics, but for a few general issues it has with the whole set of 25000 series documents:

  • The effort was to be based on taking the 9126 and 14598 sets and combining them to reduce redundancy but the result has been more proposed documents that the original two sets combined;
  • There is significant redundancy of informative material across all documents making maintenance a potential problem;
  • Most documents list most of the other documents in the set (as well as the 9126 and 14598 documents the set is to replace) as “normative references.” Such a reference is supposed to mean that the listed documents are “indispensable” to the understanding of the one in which the list appears. The existing 25000 set suggests that almost every document is “indispensable” to understanding every other one, which belies the hierarchical structure of the document set;
  • Several documents being proposed for advancement through the standards process are still missing sections or have not had an substantial changes made to them from the originals upon which they are based or have problems in that their Scope does not properly match their titles or content;
  • There are terms used without definitions and example of measures where the definition of the measure is simply a repetition of the name of the clause in which the measurement information appears.

For all of these general reasons, and some specific ones on a document-by-document basis, the U.S. has voted to disapprove all of the following documents:

  • 25001 – Quality Planning (FCD ballot)
  • 25010 – Quality model (WD circulation and CD registration)
  • 25012 – Data Quality Model (1 st CD ballot)
  • 25021 – Quality measure elements (DTR Ballot)
  • 25030 – Quality Requirements (2 nd FCD ballot)
  • 25040 – Evaluation reference model and guide (WD circulation and CD registration)

You can see that a few of these have already gotten past a few levels of balloting with the problems noted above remaining intact. Others are at or near the beginning of their formal balloting with such problems. The U.S. feels that many of these large issues should be addressed before proposing that a WD be moved forward.

A New Work Item Proposal for general industry framework for usability-related information has been proposed as the U.S. has voted to approve, offering up an editor from the U.S.

(System and Software) Life Cycle Management (TG 7)

ISO 16326 on software project management nnn

IEEE 1220 reviewed the fast tracking of this document nnn

ISO 15939 -- which covers the measurement process for software – has an NWIP proposing small modifications that would make it suitable for systems as well as software. The small number of changes to do this are because the document was originally developed with this in mind.

Process Assessment (TG 10)

ISO 15504-5 was published as an IS on March 7. This is an informative part of the standard set, containing an exemplar Process Assessment Model (PAM) based on all the processes in ISO 12207. Its publication marks the completion of the original WG10 work that began with over 5 years of effort to produce a set of nine Technical Reports. Over the past 5 years, the effort has been to take those nine and, from them, produce a set of 5 full ISs, which has finally been accomplished with the publication of this fifth part.

There are now two further documents that have been proposed for the 15504 set:

  • ISO 15504-6: a systems (ISO 15288) based exemplar PAM;
  • ISO 15504-7: an organizational maturity assessment scale allowing a typical 15504 process capability assessment profile, using a continuous model, to be turned into a staged model showing organizational maturity.

The U.S. has been opposed to the first of these on the same grounds as its opposition to 15504-5: that such documents are maintenance problems since, to be claimed as complete, they must track changes to the documents upon which they are based (ISO 12207 and, now, ISO 15288). These documents are undergoing change at the moment, for example. Also, to be an “example” of a PAM, the U.S. has always argued, does not require that every process be modeled and would reduce the maintenance exposure.

As to the organizational scale, the U.S. is waiting to see what will be proposed as being able to provide an organizational maturity conversion would be very useful, but it is not clear, given the scope of processes not restricted to software to which 15504 can be applied, what this might look like.

Functional Size Measurement (TG 12)

The major news from this Working Group is that is may be disbanded at the upcoming Plenary in May in Bangkok since all of its work program has been completed as recent votes have agreed not to advance two documents from TR to IS status. There have been no future meetings of this WG planned beyond the Plenary, which is a sign that the Convenor of the WG supports its dissolution.

[When a WG completes its originally chartered program of work, it is supposed to be disbanded as a WG is created to fulfill certain goals as specific in NWIPs, which lead to its formation. A number of WGs, however, as their initial work program comes to an end, generate further NWIPs and maintain their existence. WG10 and WG6 are examples of this. WG7’s life continues to be extended because the work to harmonize software and systems was widely supported. Also, much new work (e.g., maintenance and risk management processes) and the work of other disbanded WGs, when it comes up for revision (e.g., the measurement process), has been transferred into WG7 due to an SC7 decision, not a WG7 request.]

Software and Systems Engineering Consolidated Vocabulary (TG 22)

The IEEE has developed a web-based interface to a database where the definitions from all active (and some proposed) IEEE S2ESC and JTC1/SC7 standards will be uploaded. After all definitions are in the database, an effort will begin to remove redundancies and rank them in a preferred order. However, once all the definitions have been uploaded and properly edited, the database will be opened to the public for reference and use. The plan is that anyone can use any of the definitions and reproduce the text in any way as long as credit is given for the source. This is to encourage consistent use of terms internationally.

[There were no presentations from TG 9 or Special Working Group 5.]

A General Comment About Standards Progress and Content

There is a time limit on how long a document is supposed to have to move through the process, otherwise, it can be cancelled. It is my personal opinion that many WGs take on very large issues and have many documents, which can result in an urgency to get them moving through the process with the hope that things “can be fixed” along the way. This is a dangerous way to approach things since, if there are that many issues putting the deadline for the progress at risk, it suggests to me that the material is not really ready for an international standards ballot and should not have been proposed as such.

WG10, recognizing such a problem with the initial nine documents and the nature of the new assessment scheme being proposed, targeted the initial work at a set of TRs rather than ISs. Even with this approach, it was very hard to get some of the work done (e.g., 15504-5). Other WGs have mixed TRs and ISs in various ways, but there has been some reluctance to take the TR route for others despite the difficulty in moving them along with appropriate editing and revision.

Unfortunately, if one looks at the voting history on SC7 documents, there is always a subset of votes in the Affirmative for any and all ballots on any documents almost guaranteeing that they can be approved to reach the next level. This means documents can proceed quite far into the process from WD to CD to FCD to DIS to FDIS with issues the U.S., and some other nations, feel are problematic.

On the other hand, there are many very good ideas and information in the material, which is why the U.S. often feels that the TR approach should be taken, after which the documents can then, as they did for WG10, be revised for full IS progression.

Proposed WG25 on IT Service Management

Darcie Destito of Affiliated Computer Services (ACS) gave a presentation on a proposal to form a new Working group to focus on IT Service Management, which is covered by the ISO 20000 document set. ISO 20000 is the international adoption of the British Standards Institute’s BS15000 material on this subject. Organizations around the world, including in the U.S. (e.g., ACS being one of them) have already been registered to the standard. There was some discussion regarding the applicability of this work to the SC7 program, but, as noted earlier, a number of projects have found their way to SC7 that are only indirectly related to software and system engineering and development.

SC7 Scope Changes

Two scope change issues were raised at this TAG meeting.

One was a change to the scope of SC7 itself in order to permit inclusion of the ISO 20000 series into its work program. While there was general agreement with the scope wording, the inclusion of “methods” in the scope was a concern for a couple people who feel that the SC7 program has already expanded to take in more than its fundamental software and systems goals.

The other scope issue is more unusual. A number of years ago, an entire SC addressing modeling languages was disbanded. For whatever reason, its work was moved into SC7, but the people involved had their own TAG from the prior SC. The other TAG was allowed to continue, though the work was, officially, under the SC7 TAG. This, in itself, has been a problem since the other TAG advances documents that the IEEE Administrator often does not see coming and that the SC7 TAG never really gets to review. Now the other TAG wants to change its scope and the proposed change is very broad with the wording appearing to overlap, not only other areas in SC7 but other ISO Technical Committees! There was considerable discussion of how the US TAG Chair, Mike Gayle, would need to address this significant scope change at the Bangkok Plenary. [The issue of IEEE having two TAGs covering parts of SC7 business is also being taken up within IEEE, apart from SC7, since that it a U.S.-only matter.]

Changes in TAG membership

It looked like it might have been possible that 5 new members would be voted in at this meeting. Due to some representatives being unable to make it, the only organizations that ended up being voted in were: Human Factors International, Affiliated Computer Services, and the Project Management Institute. A division of Sun Microsystems and the Department of Homeland Security were not represented and could not be voted in at this meeting. There was also an application from Cisco, but there was no representative present from them at the meeting.

[JPL, as reported at the last TAG meeting, did not withdraw. The official withdrawal letter was never submitted, so their membership remained intact.]

Next TAG Meeting(s)

The next US SC7 TAG meeting will be at the University of Washington next September 18-20.

Those interested in any of the topics mentioned above, or other standards-related issues, can send email to sduncan@computer.org.

ASQ News

 

Follow the
Software Division

Twitter  LinkedIn