Loading ...
Sorry, an error occurred while loading the content.
 

Re: [vsx-dis] More on new discoveries

Expand Messages
  • arne
    ... There is a difference between an individual researcher submitting a single star, and importing a catalog. For the former, we require supporting evidence -
    Message 1 of 13 , Aug 31, 2006
      martin_piers_nicholson wrote:
      >
      >
      > Related to the issue of what sources potential new discoveries should
      > be checked against is the issue of how much data needs to be presented
      > to support a claim for variability?
      >
      > For example MisV1101 does appear in VSX but as far as I can see from
      > the MISAO website this is only supported by a total of 6 data points
      > taken on three days:
      > http://www.aerith.net/misao/data/misv.cgi?1101
      > <http://www.aerith.net/misao/data/misv.cgi?1101>
      >
      > A possible complication is the different focus of the two images shown
      > on the website.
      >
      > So I ask the experts - how much data needs to be presented to support
      > a claim for variability?
      >
      There is a difference between an individual researcher submitting
      a single star, and importing a catalog. For the former, we require
      supporting evidence - a light curve, a period analysis, something
      that confirms variability. This is somewhat subjective, and often
      the moderator will contact the submitter to ask for additional
      information.

      For a catalog, it is taken as a whole. In the case of the NSV,
      it was taken for historical reasons. If the NSV were created today,
      it would not be in VSX as these are *suspected* variables with little
      or no supporting evidence and often very poor coordinates. The GCVS
      team is currently going through the NSV to provide proper coordinates
      and identification for all stars - I wish them luck.

      In the case of MISAO, it is a catalog of 1340 objects, something impossible
      to check on a one-by-one basis. (Note: it includes variables either
      discovered by Martin, or on some of his images.) It was started as
      a master's project for Seiichi Yoshida, contains supporting evidence
      in the form of images and light curves for each object, gives solid
      reasoning for the selection of each star
      http://www.aerith.net/misao/variable/discoverer.html
      has had numerous refereed papers on stars contained in the catalog,
      and is coauthored by well-known variable star analysts including Taichi
      Kato.

      There will be some mistakes in such large catalogs, but we spot-checked
      before agreeing to include it and found the catalog to be of high overall
      quality. The particular star you chose had at least two images at each
      epoch that agreed with one another. However, it would be a marginal case
      for inclusion in VSX if it were submitted separately, since the two basic
      epochs were from different observers and were unfiltered images. It deserves
      study by another observer to confirm its variability, and removal from
      VSX if variability is not confirmed. You are correct in pointing this
      star out, and need to submit an "additional remark" regarding it so that other
      researchers realize its questionable status.
      Arne
    • martin_piers_nicholson
      ... impossible ... I would not want colleagues to feel that I had any significant role in the vast amount of work done by the MISAO team. I think the total of
      Message 2 of 13 , Aug 31, 2006
        >
        > In the case of MISAO, it is a catalog of 1340 objects, something
        impossible
        > to check on a one-by-one basis. (Note: it includes variables either
        > discovered by Martin, or on some of his images.)

        I would not want colleagues to feel that I had any significant role in
        the vast amount of work done by the MISAO team. I think the total of
        my contribution is two variables found in one batch of images I
        supplied. The overwhelming majority of the new variables are down to
        Ken-ichi Kadota and Nobuo Ohkura.

        http://www.aerith.net/misao/offer/list.html

        I will work through their list "as and when".
      • arne
        ... That will be a valued contribution to the community. I hope that you will also consider performing some followup observations of your own - it would be
        Message 3 of 13 , Sep 1, 2006
          martin_piers_nicholson wrote:
          > >
          > > In the case of MISAO, it is a catalog of 1340 objects, something
          > impossible
          > > to check on a one-by-one basis. (Note: it includes variables either
          > > discovered by Martin, or on some of his images.)
          >
          > I would not want colleagues to feel that I had any significant role in
          > the vast amount of work done by the MISAO team. I think the total of
          > my contribution is two variables found in one batch of images I
          > supplied. The overwhelming majority of the new variables are down to
          > Ken-ichi Kadota and Nobuo Ohkura.
          >
          > http://www.aerith.net/misao/offer/list.html
          > <http://www.aerith.net/misao/offer/list.html>
          >
          > I will work through their list "as and when".
          >
          That will be a valued contribution to the community.
          I hope that you will also consider performing some followup
          observations of your own - it would be good to see your backyard
          observatory contributing data again.

          One of my real concerns are the papers either by various surveys
          or by data miners of those surveys. The refereeing process is supposed
          to catch the majority of errors in papers, but when looking at a large
          list of objects, this process can fail.

          For example, I just reviewed a paper from a GRB group that had found
          over a thousand variables in their images. Looking at the data, I would
          agree - these are variables. However, they spent half of their paper
          describing their automated method of deriving periods and classification,
          and when examining the table based on these facts, it was very obvious
          to me that the automated process was not working. Would another reviewer
          have caught this? We've rejected a couple of lists for VSX because
          inspection by the team made it clear that the false-positive rate was
          too high; some of these were refereed publications. Surveys represent
          a real challenge to proper and timely refereeing.
          Arne
        • martin_piers_nicholson
          I have looked at the first 10 entires in the MISAO database of new variable stars http://www.aerith.net/misao/data/misv.cgi?en0100 Using VSX seems to show
          Message 4 of 13 , Sep 1, 2006
            I have looked at the first 10 entires in the MISAO database of "new
            variable stars"

            http://www.aerith.net/misao/data/misv.cgi?en0100

            Using VSX seems to show that 8 are known GCVS variables
            (#1,2,3,4,5,6,8 and 10)

            #7 only has seven data points on three days does also link to NSVS
            #9 only has six data points on three days but does also link to NSVS

            I don't really see that the MISAO data adds much to the debate.
            Should these just appear in VSX named as GCVS or NSVS variables
            rather than MISAO variables?

            As people probably know I do most of my observing using remote
            facilities in New Mexico. It would be simple to write a script that
            would take weekly V band images of say 50 NSVS variables (or indeed
            50 MISAO variables) every week for a few months.

            BUT - would this be a sensible use of observing time? Would these go
            into VSX as Nicholson 1 to Nicholson 50 <grin>?
          • arne
            ... If you look at the references, you will see that these are known GCVS variables because they were discovered by MISAO and published in IBVS. ... That is
            Message 5 of 13 , Sep 1, 2006
              martin_piers_nicholson wrote:
              >
              >
              > I have looked at the first 10 entires in the MISAO database of "new
              > variable stars"
              >
              > http://www.aerith.net/misao/data/misv.cgi?en0100
              > <http://www.aerith.net/misao/data/misv.cgi?en0100>
              >
              > Using VSX seems to show that 8 are known GCVS variables
              > (#1,2,3,4,5,6,8 and 10)
              >
              If you look at the references, you will see that these are "known
              GCVS variables" because they were discovered by MISAO and published
              in IBVS.

              > #7 only has seven data points on three days does also link to NSVS
              > #9 only has six data points on three days but does also link to NSVS
              >
              > I don't really see that the MISAO data adds much to the debate.
              > Should these just appear in VSX named as GCVS or NSVS variables
              > rather than MISAO variables?
              >
              That is how they appear. Right now, VSX lists all records from all
              catalogs, so you will get multiple hits from stars that appear in
              more than one catalog. Eventually there will be a "primary record",
              where the best information from each catalog is brought together.

              > As people probably know I do most of my observing using remote
              > facilities in New Mexico. It would be simple to write a script that
              > would take weekly V band images of say 50 NSVS variables (or indeed
              > 50 MISAO variables) every week for a few months.
              >
              > BUT - would this be a sensible use of observing time? Would these go
              > into VSX as Nicholson 1 to Nicholson 50 <grin>?
              >
              They would be entered as VSX Jhhmmss.s+ddmmss
              If you published them elsewhere, then there would be a cross-id to
              Nicholson xx if that is how you listed them in the other paper, or
              to NSVS xxxx if that is how they are indexed. There is very little
              personal glory in discovering or studying variable stars; as opposed
              to "minor planet" (parentheses mine, based on the latest IAU meeting)
              studies, where the discoverer can name the body, very few stars are
              known by an individuals' name.

              What is a sensible use of observing time depends on your personal
              opinion, not anyone elses. Likewise, whether you use the nice system
              in your backyard, or RAS, is your choice. I'm just pushing you to
              restart your Daventry system too. :)
              Arne
            • ncam1999
              ... The simplest of investigations reveals MisV0001 is in the GCVS primarily because it was published in an IBVS by MISAO and the GCVS picked it up later and
              Message 6 of 13 , Sep 1, 2006
                --- In vsx-dis@yahoogroups.com, "martin_piers_nicholson"
                <martin_piers_nicholson@...> wrote:
                >
                > I have looked at the first 10 entires in the MISAO database of "new
                > variable stars"
                >
                > http://www.aerith.net/misao/data/misv.cgi?en0100
                >
                > Using VSX seems to show that 8 are known GCVS variables
                > (#1,2,3,4,5,6,8 and 10)
                >
                > #7 only has seven data points on three days does also link to NSVS
                > #9 only has six data points on three days but does also link to NSVS
                >
                > I don't really see that the MISAO data adds much to the debate.
                > Should these just appear in VSX named as GCVS or NSVS variables
                > rather than MISAO variables?

                The simplest of investigations reveals MisV0001 is in the GCVS
                primarily because it was published in an IBVS by MISAO and the GCVS
                picked it up later and formally named it, as has happened with many a
                MISAO variable

                http://www.sai.msu.su/groups/cluster/gcvs/cgi-bin/search.cgi?search=V4652+Sgr

                In this primary instance the Mira star was found from images of the
                field of the adjacent suspected cataclysmic variable V4334 Sgr via
                using PIXY SYSTEM (1) on donated images for that field.

                The GCVS reference to the study of MisV0002 is again to an IBVS by
                Yoshida et al on MISAO variables.

                VSX also echoes this information. Probably same for t'others, I can't
                be bothered to look.

                MISAO variables predate NSVS red variables in many instances, memory
                tells me up to MisV0500 were published in IBVSs, although in what form
                I remember not, and further that these are subsequently in SIMBAD, so
                the NSVS team should have found them via cross checking. In fact the
                NSVS team did not even fully cross check against the GCVS and NSV as
                at that time said catalogues were only up to date in the SAI ftp
                archive and not integrated into SIMBAD, which had a woefully out of
                date copy. MisV often came afore subsequent published NSVS J, in
                other words, and just as often the NSVS solution is no more accurate
                and many of their long period variable periodicities are between
                farcical and very bad indeed.

                The date is likely around 2003 when which survey has priority probably
                swaps, last I was actively taking note of MISAO stuff and doing cross
                indexing work therefor up to around the MisV1200s, MisV1147 and the
                potential Nova Cephei 2001 = V0709 Cephei which I think = MisV1181 was
                about the last ones I remember. There was no NSVS data to play with
                at that time. Amateurs using visual techniques provided a lot of the
                full lightcurve for the semilegendary MisV1147, plus some allsky CCD
                several passband photometry from folk like Ondrej Pejcha. It was
                quite a little international get together on the quiet was that one.


                It is within the context of these matters that VSX can do work if folk
                experienced in such work are willing to cross correlate multiple
                entries for single objects, show the connection with reference to
                evidence, literature based or new observation based, and tighten up
                the reportage system of the database. That is what it is about. VSX
                makes the MISAO and GCVS and NSVS data available for the tidying up of
                the case of what the true situation is re the MISAO _discoveries_.

                As is, the MISAO list was imported en masse and therefore duplicates
                other entries, as many were published by MISAO, or noted to vsnet
                lists as was, and subsequently were formally adopted into the GCVS,
                which too lives in VSX.

                I'll repeat that a third time. MisVnnnn objects are often in GCVS
                because they were discovered as new by MISAOchan, checked against then
                current catalogues for newness by her alter ego Pixy Misa, and double
                checked for knownness by Kato Taichi and independently for some by
                John Greaves, although neither were formal "MISAO members".

                The NSVS database permits folk to solve such objects if they so wish,
                indeed Yoshida Seiichi himself typed and periodified several long
                known red MISAO objects using the extra data thereby made available,
                sometimes in tandem with new CCD observations by Japanese observers.

                Solution is the emphasis. The current situation reminds me that I
                advised strongly against full MISAO list import into VSX as there are
                _some_ gray area objects whilst the more certain objects are fully
                published in IBVSs (sadly IBVS data still ain't in VSX for some
                unfathomable reason) and/or already integrated into GCVS and later
                namelists. I took the view that misinterpretations re VSX submission
                could ensue based on one or two, and only a handful mind you, of cases
                carefully picked out of MISAO for agenda based illustration by those
                desirous of so doing.

                A list, no matter of what longevity, that contains some objects only
                appearing in that list on a webpage being imported into VSX could
                potentially be an unfortunate precedent, was another view I took.

                However, Seiichi is a devoted and rigorous individual, and took every
                effort to ensure his findings were new afore calling them as such.

                Remember again that the vast majority of the MisV objects appeared
                afore the current surveys and lists began to make their data public.

                > As people probably know I do most of my observing using remote
                > facilities in New Mexico. It would be simple to write a script that
                > would take weekly V band images of say 50 NSVS variables (or indeed
                > 50 MISAO variables) every week for a few months.
                >
                > BUT - would this be a sensible use of observing time? Would these go
                > into VSX as Nicholson 1 to Nicholson 50 <grin>?

                Why not Nicholson 1 to twenty million? Yes, they'd likely go into VSX
                as Nicholson 1 to 20,000,000 provided they were shown variable, shown
                truly to be new and/or shown to be properly and fully analysed and
                solved if their existance in other surveys, such as ASAS3 and NSVS,
                was under a false analysis/variability typing, as perforce they would
                be new solutions for either known or new objects. However, in the
                case of preknown but incorrectly classed objects a note should be made
                of their preknownness. If only for bibliographic reference... ...the
                old paper trail, don'tja know.

                The nature of the MISAO system is somewhat historical, it preceded the
                current datamineable (more properly, datatrawlable) surveys and was
                based on processing of donated CCD observations taken by amateurs.

                It was set up to find new asteroids by happenstance in spare images,
                something soon made irrelevant by LINEAR, and they hoped one day to
                find a nova or two, and they more or less did find one in the end.

                Whether such a scheme is strictly relevant in the modern age is
                unclear.

                A system likely more relevant to the current way of things is likely
                Vello Tabur's survey of the southern sky with a simple system,
                something he unfortunately never wrote up, coz it were impressive, and
                it found truly new variables, and solved their nature, and not all of
                those were found or included in subsequent ASAS3 releases either.

                John
              • Sebastian Otero
                ... Actually if the observations are for Misao (or any other known) variables they will enter VSX as revisions of the already existing stars so if they already
                Message 7 of 13 , Sep 1, 2006
                  >> BUT - would this be a sensible use of observing time? Would these go
                  >> into VSX as Nicholson 1 to Nicholson 50 <grin>?
                  >>
                  > They would be entered as VSX Jhhmmss.s+ddmmss
                  > If you published them elsewhere, then there would be a cross-id to
                  > Nicholson xx if that is how you listed them in the other paper, or
                  > to NSVS xxxx if that is how they are indexed.

                  Actually if the observations are for Misao (or any other known) variables
                  they will enter VSX as revisions of the already existing stars so if they
                  already have a designation I don't see why a new name should be added just
                  for personal glory satisfaction...
                  If every variable star researcher had that purpose in mind when it is
                  analysing a new or unsolved variable, the nomenclature would be chaotic.

                  Cheers,
                  Sebastian.
                • martin_piers_nicholson
                  As a way to greatly speed the process. Ask the experts to generate a list of the MISAO variables not currently linked to a GCVS entry. Better still do this
                  Message 8 of 13 , Sep 2, 2006
                    As a way to greatly speed the process.

                    Ask the experts to generate a list of the MISAO variables not
                    currently linked to a GCVS entry.

                    Better still do this concentrating on those south of -30.

                    You will find most, if not all, of these are very questionable.


                    MISAO 50-100 but not in GCVS

                    Very questionable
                    52
                    56
                    59
                    64
                    77
                    78
                    79
                    80
                    83
                    89
                    97
                    98

                    Questionable
                    58
                    68
                    90
                    94
                    96
                  • Arne Henden
                    Obviously, I m a little confused here. For example, #52 may have few data points, but it has nearly 2 magnitudes variation, and the images are pretty clear as
                    Message 9 of 13 , Sep 2, 2006
                      Obviously, I'm a little confused here. For example, #52 may have few
                      data points,
                      but it has nearly 2 magnitudes variation, and the images are pretty clear as to
                      which star is variable and by how much. They are even from the same observer.
                      Are you saying this is not a variable? Or are you saying that you
                      want a minimum
                      amount of information, such as a complete light curve and period
                      analysis, before
                      any star is considered variable? It certainly meets your amplitude > 2-3 sigma
                      criteria. So what do you mean by very questionable? I'm not debating with you,
                      just trying to understand your concern. 64 is similar, and is linked
                      to V1174 Sco.
                      I didn't look further into your list.
                      Arne

                      On 9/2/06, martin_piers_nicholson <martin_piers_nicholson@...> wrote:
                      > As a way to greatly speed the process.
                      >
                      > Ask the experts to generate a list of the MISAO variables not
                      > currently linked to a GCVS entry.
                      >
                      > Better still do this concentrating on those south of -30.
                      >
                      > You will find most, if not all, of these are very questionable.
                      >
                      >
                      > MISAO 50-100 but not in GCVS
                      >
                      > Very questionable
                      > 52
                      > 56
                      > 59
                      > 64
                      >
                    • martin_piers_nicholson
                      Take Misao 98 as an example. BAD NEWS Only 2 data points, March and June 1999. However to my eye quite a few other stars in the field look different as well.
                      Message 10 of 13 , Sep 2, 2006
                        Take Misao 98 as an example.

                        BAD NEWS
                        Only 2 data points, March and June 1999.

                        However to my eye quite a few other stars in the field look
                        different as well.

                        GCVS didn't include it although it appears in the same issue of IBVS
                        as many MISAO stars that were accepted.


                        GOOD NEWS
                        They differ by a quoted 1.1 magnitudes and the photos clearly show
                        the indicated star has changed in brightness.



                        What I would like is clear guidance on what constitutes proof of
                        sufficient quality and quantity to qualify for inclusion in VSX. If
                        I were to invest time in going through the material it would have to
                        be on the basis of having this framework in place first. (Refer to
                        forthcoming email on priorities in this context)
                      • kawai.sasami
                        ... You have evidently not read up on or learnt about or taken note of the pixy system and its subsequent revised form the pixy2 system. If you had you would
                        Message 11 of 13 , Sep 3, 2006
                          --- In vsx-dis@yahoogroups.com, "martin_piers_nicholson"
                          <martin_piers_nicholson@...> wrote:
                          >
                          > Take Misao 98 as an example.
                          >
                          > BAD NEWS
                          > Only 2 data points, March and June 1999.
                          >
                          > However to my eye quite a few other stars in the field look
                          > different as well.

                          You have evidently not read up on or learnt about or taken note of the
                          pixy system and its subsequent revised form the pixy2 system.

                          If you had you would know that differences between brightnesses
                          between discordant images are allowed for, correction can be applied
                          given the reasonable assumption that the vast majority of the field
                          stars are constant. Once this form of calibration has been applied
                          any remaining difference in magnitude found for any particular star
                          has a likelihood of being variability, the large the amplitude, the
                          higher the likelihood.

                          As most images were unfiltered highish amplitude red LPVs were
                          predominantly detected in the MISAO system.

                          Incidentally, I am not part of and never was part of MISAO and I shall
                          note that separately.

                          > GCVS didn't include it although it appears in the same issue of IBVS
                          > as many MISAO stars that were accepted.
                          >
                          >
                          > GOOD NEWS
                          > They differ by a quoted 1.1 magnitudes and the photos clearly show
                          > the indicated star has changed in brightness.

                          they were measured, reduced, to show this. Not guessed by eyeball.
                          Granted the photometric calibration is against USNO A2.0 red mags, so
                          the magnitudes are purely MISAOing, but that's the way it was.

                          The original intent was to find incidental asteroids in spare images.
                          PIXY and any presumable precursor thus had to match any and all
                          images to the USNO A2.0 catalogue to find new objects that shouldn't
                          be in the field. Some level of magnitude estimate is also useful for
                          such endeavours, and A2.0 at the time, and still now to some extent,
                          is the only readily available bulk system allowing stars to fainter
                          mags than most amateur imaging. ie Seiichi has a copy. Although I
                          vaguely remember PIXY2 could fetch other data off of the net later on.

                          All form memory, all stuff I learnt by bothering to read the MISAO
                          website in the past.

                          John
                        Your message has been successfully submitted and would be delivered to recipients shortly.