NOAO Telescope Proposal Process Workshop
10-11 August 1998
Representatives of several observatories met to discuss common issues and goals related to telescope proposals for astronomical telescopes. The goal of the Proposal Process Workshop was to develop a shared understanding among observatories, including the National Gemini Offices and the Gemini Project, of the requirements and procedures for telescope proposals, and to encourage cooperation among the national observatories of the partner countries, STScI, and other institutions faced with similar issues and concerns related to the telescope proposal process.
Through discussions at the Workshop, a broad consensus emerged. First, the proposal process should include an option for web-based proposal preparation and submission that does not require the distribution of software. For now this was the best option for the US community, but we will need to continually re-evaluate the situation as new tools and methods become available.
Second, that the process should be kept as simple as possible for users, with required information kept to a minimum, while allowing investigators to present their case in a suitable manner. The process should also minimize the effort required at each observatory to support TAC evaluation and observing. The process should allow integration of proposals for the variety of telescope time available through each observatory.
Actions resulting from the Workshop include the distribution of the NOAO Web proposal system to several observatories, and the formation of two working sub-groups to develop a set of LaTeX/XML tags to be shared among observatories and to develop guidelines for layout and organization of Web-based tools for proposal submission that will allow the sharing of tools among observatories.
Representatives of several observatories, including NOAO, Gemini, STScI, McDonald Observatory and the HET, the MMT Observatory, the CFHT, the National Research Council of Canada, the AAO, and JACH, met together in Tucson to discuss common issues and goals related to telescope proposals for astronomical telescopes. The goal of the Proposal Process Workshop was to develop a shared understanding among observatories, including the National Gemini Offices and the Gemini Project, of the requirements and procedures for telescope proposals, and to encourage cooperation among the national observatories of the partner countries, STScI, and other institutions faced with similar issues and concerns related to the telescope proposal process. A list of participants, the meeting agenda, and links to proposal related information available through many observatory web sites is posted at the Workshop website at http://www.noao.edu/scope/tpp_workshop/.
Broad issues and goals were explored during the first morning, and the first afternoon was used to investigate the content of proposals to identify the complete set of information necessary to include in a telescope proposal. On the second morning we discussed the proposal process to identify in some detail the scientific and procedural requirements for the different phases, from preparation through submission, merit review, technical review, and scheduling. On the final afternoon we looked more deeply into specific implementation issues, with the goal of identifying a clear path to providing a set of tools for investigators and observatory staff to manage the proposal process.
Boroson described the changing role of NOAO from building and operating telescopes on its own sites to supporting community access to a wide range of facilities, not all of which are operated by NOAO. NOAO wants to provide the community with an integrated system to enable astronomers to carry out scientific programs in new ways across a range of facilities. NOAO has already implemented a single proposal form and submission process for both CTIO and KPNO. In March, 1999, NOAO expects to merge the KPNO and CTIO TAC processes and to accept for the first time proposals for community access time for the HET and the MMT. NOAO will be ready to accept its first proposals for the Gemini telescopes in Sept. 1999.
Puxley presented the vision of the Gemini Project for proposals for time on the Gemini Project. Programs approved and ranked by the individual partner countries for both queue and classically scheduled time will be forwarded to Gemini. A draft schedule will be prepared by the operations staff. That schedule will be reviewed by the Gemini International TAC, and the schedule will be finalized. Because the Gemini telescopes will be operated half in queue mode and half classically, and because the performance of the telescope depends critically on environmental conditions (seeing, water vapor), investigators must supply detailed observing constraints for carrying out programs. The feasibility of programs using the AO system as well as most of those requiring the on-instrument wavefront sensor must be established by demonstrating that a suitable guide star is available; for other programs, pre-selection of guide stars is not as critical. First light for Gemini North is expected in December 1998, and the Gemini Science Committee has called for the first TAC process to be planned for September 1999. Proposal deadlines will be current deadlines at NOAO and other observatories.
Miller and Blacker summarized the experience of STScI after 13 years of proposals and 8 years of operations for the Hubble Space Telescope. In their most recent complete cycle, almost 1300 proposals were submitted. Over the course of 7 completed proposal cycles, STScI staff has simplified more and more the information which must be provided within the Phase I proposal. The have succeeded in reducing the time from submission to notification from more than 6 months to 3.5 months. The presentation covered in-house proposal processing, the review process, conflicts of interest, the policy on duplication of observations, technical reviewing, investigator notification, staffing, and future plans. The lessons learned from the experience at STScI were several.
Jannuzi lead a discussion of the role of Phase I and Phase II in the context of telescope proposals. He suggested that a good definition of a Phase I proposal for the observer would be: A proposal includes a detailed description of the scientific goals and experiment as well as enough information about the proposed observations (the experimental design) that the feasibility can be evaluated. This might included a list of targets, instrument requirements, and observing modes, but probably would not include all of the information that would be needed to carry out the observations. Phase II for the proposer would involved detailed plans for the observations. For all programs to be carried out in queue mode, this would mean providing ALL the information necessary for the observatory to perform the observations on the observer's behalf.
Contents of a proposal
Barnes summarized what information is requested on existing proposal forms in the areas of general information, text and figures, investigator information, observing run information and constraints, and instrument specifications. She began with a definition of what a good proposal is.
A good proposal is one that allows the investigator to state
clearly his/her scientific goals while supplying in a concise
manner the requirements expected from the site to accomplish
these goals. The information provided through the proposal
must be sufficient for
Useful ideas emerging from this discussion included the following:
Cornell continued the discussion of the content of telescope proposals, considering target and guidestar information, environmental requirements, and exposure calculations and observing time estimates. He highlighted the differences between classical and modern observing. Software may now exist to estimate the required S/N in detail. Proposals often include detailed technical information about the observations, and observations must be planned in advance of the observing run, rather than in real time. The PI knows best what is needed to get the science done, but may have difficulty translating his/her experience to quantitative estimates of the requirements (e.g. percent water vapor or sky brightness). The traditional constraints are not always applicable (e.g. lunar phase vs. sky brightness). Discussion of these issues led to two guiding principles: 1) the PI does not want to specify any more information than he/she has to (and we especially do not want to overspecify constraints for queue observations); 2) we should allow the PI to provide constraints in familiar units, and the site translate those units into the quantitative units needed to schedule programs. The following comments concern detailed Phase I information.
Target lists - these are part of the science case, and at least an RA range should be specified. Guide stars - not typically needed or required. - required for some (but not all) Gemini modes Minimum useful time - better if provided by PI than by the TAC Lunar phase - useful shorthand for sky brightness - fine for classical mode - translate to surface brightness for queue mode Seeing - the PI wants to provide this in arc seconds - the PI and the TAC need to be able to relate a specification in arc seconds to the percentage of time this condition is met. - seeing specifications demonstrate the need for good documentation Water vapor - changes absorption in the IR a lot - show this to the PI with plots in the documentation Transparency - the site needs to know what the PI can tolerate - how can this be parameterized?If guide star lists or finder charts can be generated automatically, do so. Specify exposure times for nominal or expected conditions. Are these to be scaled by the observer at the telescope (for HET, yes).
The Proposal Process
Crabtree addressed issues related to the process of preparation and submission of telescope proposals. The critical issue here is whether proposals should be prepared and submitted over a web-based system or using distributed, standalone software. The advantages of the former are a) observers are always using current software, b) there is no need to download and install software. Disadvantages include the slowness of a network link for some users, the difficulty of using complex, interactive graphics over the Web, and the need to be connected to a network to prepare a proposal. The advantages of standalone software are a) more complex interactive graphics interfaces can be implemented and b) users need not be tied to a network to prepare proposals. The primary disadvantage is user support on a variety of platforms. Participants at the workshop held both points of view, but there was a general consensus that at a minimum, a web-based system should be provided as an option for users.
Good image quality estimators and exposure time calculators are much valued, both by the community and by the observatory sites. A ``Phase 0'' proposal tool would be useful to guide proposers in the selection of telescopes and instruments for carrying out their projects. A ``units translator'' would be a useful accessory to proposal tools to allow proposers to convert unfamiliar units. We should also have available a simple proposal process for short requests with less complex scientific justifications.
Glazebrook reviewed tools for proposal processing following submission, including database importation, software tools, support for Telescope Allocation Committees, evaluation of schedulability, technical evaluation, and scheduling. He noted the advantages of LaTeX, which is widely used: it is familiar, it is used by the journals, and it produces a nice hard copy. Its disadvantages are its flexibility, which can be abused by proposers, the difficulty of parsing information with embedded LaTeX commands, and a preference of some users for more modern text formatting tools. Glazebrook proposed a ``split'' proposal following a model implemented at the AAO. Proposers use a LaTeX file to submit the scientific case, but a web tool to provide detailed instrument information to obtain better accuracy. He noted that both Perl and Java are useful software tools for processing proposals, and that most observatories are using Perl and Procmail. New approaches for TAC support are being tried or studied at various observatories: electronic grade submission, electronic refereeing, rolling submissions. No tools seem to be available yet to write schedules, but tools to assist in writing schedules are common. Everyone wants to reduce the workload associated with technical review, and the favored method seems to be to be by adopting proposal preparation tools that enable the proposers to improve the quality of the information in the proposal. For this method to be reliable may require some sort of validation that the proper tools have been used.
Puxley described an implementation plan for proposal submission that the Gemini Project was developing in collaboration with some UK-affiliated observatories, including the AAO and JACH (JCMT and UKIRT). The plan involves using the Gemini Phase I tool which would be distributed to UK investigators via Starlink for preparation of proposals. Proposals would be submitted by investigators as XML files, including attached text and figures for the scientific case. Puxley stressed the value of a common format for encoding proposals, and noted our common interest in doing so. One goal of the submission process is to minimize effort for both the users and the observatory. This may mean moving to a Web-based system, and he hoped Gemini could move toward this goal.
Davidge described the experience of HIA in handling proposals from Canadian astronomers, and raised several issues for consideration These included concerns about confidentiality, multi-telescope proposals, language flexibility, and multi-national coordination. This last especially may drive closer coordination as international teams of investigators propose through their own national processes to carry out parts of a collaborative program. He noted that the NOAO Access database for proposals may be of use to other observatories but it will need to accommodate national differences. The need to set standards for sharing information among TAC's will be driven by Gemini. XML may be the mark-up language of choice in the future, and we should move in that direction together. For the meantime a number of more immediate issues regarding LaTeX, parsing, and translations need to be addressed.
Putting it all together
Attention then turned to a model of the telescope submission process that that might serve the needs of the several observatories participating in the Workshop. The model must accommodate several needs of the user community, the TAC, and the observatories. These include:
Boroson sketched how such a system might look (with apologies for slight modifications for clarity!):
Web proposal Parser ----> Database ---> XML form / for Gem \ / \ / \ Proposal queue Checker -> or / Processor / \ / \ / \ Latex proposal PostScript/Printing formDiscussion then turned to actions that we can take to facilitate the development of shared tools and standardization.