Structure of an ALMA Program Committee

 

 

Background

The ALMA Board has expressed interest in investigating the implications 
of a single Program Review Committee (PRC). The structure of such a 
committee is open to discussion. The Board has asked the regional 
Project Scientists to prepare a short report with some details of the 
set up of a single PRC.

 

As an appendix, the relevant section of the ASAC report of 2004 October 
is included. The ASAC was of the opinion that a single PRC is very much 
more useful than separate committees.

 

Operations Plan

The Operations Plan of ALMA, pages 46, states that ' All proposal and 
observation preparation material shall flow from external users to the 
central archive via...the Internet. This information shall then be 
replicated immediately to the OSF and ARC archive nodes.' That is, 
observing proposals are sent to JAO, and JAO distributes these to the 
local ARCs, who then process these further. In the case of a single PRC, 
one of the ARCs might handle all of the Phase I and II processing of  
the proposals, or a number of ARCs might share this work.

 

Detailed Implications

 

For a single PRC, a number of parameters need to be specified.

 

First is the frequency of the meetings. These might be either once or 
twice per year.

 

Second, there are two extremes for a PRC. These are either a 
face-to-face meeting, or a via email. In both schemes, ther would have 
to be a PRC secretariat to handle the details of scheduling, 
distribution of propposals and collection of referees grades. With both 
schemes there would have to be a Time Allocation Committee (TAC) that 
makes final decisions on proposal rankings. For a face-to-face PRC with 
sub-panels, this would be the sub-panel chairs and a TAC. For an email 
system, this would be a panel of experts (presumably at JAO) to decide 
which proposals are technically feasable and make other fine-tuning 
adjustments. There could (of course) be situations between these 
extremes. One would be email inputs to a set of sub-panels. The number 
of persons in a face-to-face could be smaller, and these person could 
concentrate on proposals for which the email results show a large scatter.

 

Third, any committee would have to consist of groups of experts for 
research areas. There are two examples that come to mind, namely ESO and 
Space Telescope Institute.  At ESO, there are 4 sub-panels. Each of 
these is divided into a set of topics. The specialites are given in the 
following with the number of topics in parentheses. These areas are: (1) 
Cosmology (8 topics),  (2) Galaxies and Galactic Nuclei (9 topics), (3) 
Interstellar Medium, Star Formation and Planetary Systems (8 topics), 
and (4) Stellar Evolution (11 topics).  After each sub-panel ranks 
proposals within each category, a meeting of the chairs of sub-panels 
would rank all of the proposals. Another example is the Hubble Space 
Telescope (HST) PRC. This has 1 meeting per year, with a face-to-face 
meeting held near the Space Telescope Institute. For the HST Cycle 15 
there are 11 panels that cover wide areas. The total number of members 
for the HST PRC is 109 members.  After the panels meet, the panel chairs 
and a few at-large members meet at a Time Allocation Committee (TAC).

 

 

Arranging to bring a large group of referees together for a face-to-face 
is a challenge.

Face-to-face meetings are much more difficult for 2 meetings per year. 
Having face-to-face meetings would help to remove any misunderstandings 
about the content of proposals. However  at ESO such meetings require a 
week of time. On the other hand, the email form of a PRC places more 
power in the hands of the TAC (presumably at JAO). Also, 
misunderstandings are more difficult to eliminate.

 

The makeup of the sub-panels is the third point. The division here is 
taken from the ESO set up. There could be finer divisions of categories, 
but then there are more panel chairs and members. This adds complexity 
and cost to this part of the project.

 

Beyond these high level considerations are more down to earth, practical 
issues. However the first goal should be to outline the high level 
constraints and only then delve into the fine details.

 

The number of PRC members is an important issue.


Dear Tom and Al

A few comments re TACs.

I think that the main difference to highlight is that between the NRAO/VLA
model, where the scheduling committee is (mostly?) internal and makes a decision
based on reports from external referees submitted by email and (e.g.)
Chandra/HST, where fully representative committees with many sub-panels meet
face-to-face.

The latter model is clearly much more expensive to run. My impression is that
the VLA model is liked by users (and by referees, as they do not have to
travel). Two important factors are that the scheduling committee has developed a
reputation for insight and impartiality (due mainly to Barry's influence, I
suspect) in the face of occasionally daft referee reports. But I wonder whether
this model would be found acceptable if the oversubscription rate were much
larger (as it is for HST) or if no Barry figure could be found.

Differences between HST/Chandra/ESO OPC (and other systems), it seems to me, are
much subtler - they all involve numbers of specialist sub-panels of various
sizes and lengthy f2f meetings.

Cheers

Robert