[Date Prev][Date Next] [Chronological]

skiLS summary/pics 

(A couple of people that were not at the Mt. Wash meeting asked how it
went, so...)
 
The Mt.Wash meeting was exceptionally good, by any standard.. It was
informative, highly professional, and fun. We had a completely full
program of extremely talks that would (and have) been given at major
astronomy meetings, covering the entire range of activities in the
survey. There followed a number of frank discussions about how
students and PDFs saw their opportunities in this activity. A number
of important ideas for everything from the design of the survey to
novel scientific interests emerged. Since we were together about 16
hours of every day everyone had a chance to interact with virtually
everyone else. I could go on to editorialize, but talking to someone
who was there would probably be best. We will have something similar
again, although the format will evolve with the survey.
 
Some rather poor quality webcam photos that don't convey the energy
level an were taken at random (less busy) moments are posted at:
 http://manaslu.astro.utoronto.ca/~carlberg/mega/skils/
(This whole site will undergo a revolution soon)
 
The following recommendations were discussed in detail during the
final session. There has been some subsequent discussion.
 
Field Locations
1) endorse Groth strip for one Ultra-deep/Sne
2) encourage further overlap with SWIRE survey
3) technical issue that 6x6 squares may be too large to be optimal
        for control of sample variance
        (much further thinking on this, with suggestions of an
        8x8 XMM field, possibly a few intermediate fields, then
        a sparse sampled region)
 
Calibration
1) All science will benefit from 0.01 mag systematic error limits
        on all fields.
2) sne data requires 0.01 maq systematic, 0.02 mag random, with
        goals of 0.005 mag systematic and 0.01 amg random across
        all fields
3) An expert group on calibration should be established, at least in
        Canada, and ideally as a joint program between Canada, France
        and CFHT.
 
 
Proprietary Time

The scientific impulse to openly share facility data is deeply rooted
in astronomy.  The Canadian community shares this view but has
expressed nearly universal concern that 13 months from time
of observation is too short. Arguments include:
 
0) various aspects of as yet unproven science quality control, ranging
           from  chunky   datasets  (missing  filters,   fields,  poor
           calibrations at 13 months)  which give a misleading view of
           quality  of  the survey,  little  time  to perform  serious
           science  analysis  as  a  reliability check,  and  possible
           delivery issues.
 
1) The CFHTLS will be the first ever, large, deep, uniform imaging
        survey. Many of the fields overlap with locations of very deep
        surveys being undertaken with other instruments.  Furthermore,
        the Elixir data is of a quality that is entirely sufficient
        for many scientific purposes with little further
        processing. The basic knowledge and capacity to deal with the
        uniform high quality CFHTLS data is becoming widely
        distributed in the astronomical community.  Whenever they are
        released the data are unlikely to languish.
 
2) The first few years are likely to create somewhat fragmentary
        datasets due to weather and other problems that will best left
        for publication for years two or three at which time the
        proprietary time will give a lead of only one year.
 
3) Undue haste in publication can be extremely counterproductive.
 
4) There is no advantage to the CFHTLS science community for such a
        short proprietary time.
 
5) Such a short proprietary time diminishes the possibilities of
        collaborative leverage of CFHTLS to gain other data,
        particularly given the somewhat slow, complex and diffuse
        structure for approval of collaborations.
 
6) Students are extremely concerned about being scooped. Rather than
        encourage people to concentrate on CFHTLS, a short proprietary
        time will in some cases cause people to concentrate on
        acquiring proprietary data at other telescopes which will give
        them control over their research science even if the CFHTLS
        results are published elsewhere. A longer proprietary time
        would allow a more balanced approach. Within the Canadian and
        entire CFHTLS commununity the CRO grant will be used to
        encourage open communication of research work and to help
        redress any serious incidents of student trampling.
 
7) Attracting science PDFs to a three year position where at least two
        of the three (or 4 of 5) years of the data will be openly
        available gives Canadian (or French) positions very little
        competitive advantage.
 
8) The 13 months is too short on the basis that most Canadian and
   French astronomers requesting followup time on a national facility
   need to apply through the TAC process to be awarded time two
   semesters after the discovery when the field is again
   visible. Someone with more flexible access to a telescope could
   make the same discovery in the database a year later, and, if
   appropriate scheduled, could scooop the person awarded time through
   the TAC process.

Based on the PDF timescale of 3 years, a feeling that the core work of
most PhD projects is done within about 3 years, and that it would be
important to begin the release of CFHTLS data while the survey is
ongoing, it was generally felt that three years would be an ideal
proprietary time. Two years would be an improvement but would still be
very short. Five years would exceed the active phase of the
survey. Although there are advantages of a short proprietary time for
recognition of the important work done at the CFHT Corporation itself,
the instrument builders and the data centres, these considerations
should not prevail over the scientific motivations which brought these
into existence.


Back to the Mailing Lists page