The composition of a program committee is one of the most important factors in determining the quality of a conference and the quality of the papers it accepts. This in turn reflects on the long-term prestige of the conference and ultimately on the impact it is able to have. The ICSME Steering Committee has thus adopted a set of guidelines for use by ICSME Program Chairs in assembling a program committee.

Program Chairs are asked to submit their list of prospective PC members to the Steering Committee Chair at least one month prior to the desired date for sending out invitations to the PC. In any cases in which prospective PC members, or the list of members as a whole, does not meet these guidelines, the Program Chairs should indicate this, and provide arguments for these cases. The Steering Committee will consider, on a case by case basis, such arguments, as it evaluates the PC list as a whole. The Steering Committee reserves the right to require changes in PC composition based on these guidelines.

The application of this process and these guidelines commenced with the PC for ICSM 2004. To facilitate the phasing in of the guidelines, however, the Steering Committee allowed extra flexibility in the application of the rules for ICSM 2004 and ICSM 2005. Guidelines 3a and 3b have been revised in 2014 and will apply to ICSME 2015 onwards. The document has been further modified to reflect the decision to change the conference name from the IEEE International Conference on Soft- ware Maintenance (ICSM) to the IEEE International Conference on Software Maintenance and Evolution (ICSME), from 2014 onwards.

  1. 1. ICSME reviewers need to be able to judge whether research submitted to ICSME is techically sound, provides a contribution to the field, and has novelty with respect to previous work. To ensure that reviewers have this qualification, we require that to serve on the ICSME PC, a person must have had an active and documented role in the field of software maintenance and evolution in the preceding 5 years. Although we could here define a metric with which to assess qualifications, we prefer not to do so in order to allow flexibility. However, on prospective PC members whose qualifications in this regard are not clear by some obvious metric such as relevant publications, we request that the PC Chairs briefly summarize those members qualifications when presenting their list to the Steering Committee.

  2. 2. ICSME reviewers need to have familiarity with ICSME itself, and with the review process, from the point of view of an author. To ensure this, we require that first-time PC members must have previously authored or co-authored at least one paper that has been accepted to ICSME.

  3. 3. Program committees require continuity to ensure that conference goals can continue to be met. It is also important, however, that PCs make room for new researchers, and that Program Chairs do not feel obliged to retain, for historical reasons, committee members who do not fully participate in the process of preparing high-quality reviews. Thus, we require that PCs explicitly incorporate a process of rotating members on and off of the PC, as follows:
    • a. No PC member shall serve on more than three consecutive PCs, following which they must be omitted from the PC for at least one year. Acting as a conference general chair is not considered as being a member of the program committee. Acting as a program committee (co-)chair is considered as being a member of the program committee. However, a person that has been on the PC for three consecutive years, can act as the PC chair in the fourth year.
    • b. On each main (research) track PC, at least 30% of the members must be new with respect to the preceding year's main track PC. All other tracks should encourage high turnover of the PC members.
    4. </li>

  4. 4. To help ensure the success of future ICSMEs, the Program Chairs and General Chair associated with a future ICSME should be on the PC for ICSME for the year preceding that ICSME. (In this instance it is permissible to violate guideline 3a.)

  5. 5. Given the need to provide effective reviews in each of the many sub-areas of research covered by ICSME, Program Chairs must ensure that the PC includes members whose areas of expertise sufficiently cover those areas of research.

  6. 6. Given the desire to continue to project ICSME as an international and inclusive conference, Program Chairs should make every effort to achieve diversity on the PC with respect to gender, geographic distribution, ex- perience, and industry versus academic experience.

  7. 7. The quality of a conferences reviews is central to the view authors have of the conference, and to the conferences subsequent prestige. This in turn affects the conferences ability to attract good work and have an impact. Therefore, Program Chairs should make every effort to invite PC members who are expected to and agree to abide by the following proscriptions as a condition of serving on the PC.
    • a. PC members are responsible for their own reviews. Although it is acceptable to obtain co-reviewers such as students, who may provide relevant expertise, it is not the student who is on the PC, it is the PC member, and ultimately that member is responsible for the review and must be personally able to argue for or against the merits of the paper.
    • b. PC members must play an active role in helping authors improve their papers. Reviews should contain details sufficient to support their conclusions, and reviews should be constructive, offering comments on how papers can be improved.

    To ensure that authors are aware of and agree with these proscriptions, Program Chairs should include them prominently in any invitation sent out to prospective PC members, stating there that acceptance of the in- vitation implies agreement with the proscriptions.

  8. 8. To determine the size of a PC, calculate an estimate on paper submissions E given the submission numbers from the preceding two ICSMEs, and determine the number of PC members necessary to handle E papers, consistent with having 3 reviewers per paper, and a reviewing load of between 7 and 10 papers per member.