[QUANTIT] Use-cases, role in larger scheme (Was: Re: [QUANTITY] Quantity "arguments")

[Brian Thomas]

On Friday 14 November 2003 02:46 pm, Doug Tody wrote:
> On Fri, 14 Nov 2003, Steve Lowe wrote:
> > I think what is missing is requirements.
>
> Component-level requirements are needed, but we need to back up one
> step further and define an architecture which lays out what the data
> model components (such as quantity) are, and what their role is in some
> larger scheme. 

	Ok, towards that end, I present 2 use-cases, with starting thoughts on
how these produce requirements which indicate a need for a quantity.

The following 2 use-cases were taken from the NVO SWG list.

I have put comments following each use-case indicating as to how this
pertains to quantity.


Regards,

-b.t.


2-use cases for "quantity", based on search scenarios:

-----------------------------------------
1. GRB follow-up service

- Upon a receipt of a GRB detection message (from GCN, SAX, HETE,
Swift...):

- Grab the image and catalog data from all linked archives covering the GRB
error box

- Cross-match any sources in this set

- Identify possible permanent-but-variable sources (i.e., not 1-time
transients) in this area from multiple observations at the same (+-)
wavelengths (e.g., optical, others if available)

- Prepare the finding charts with all the sources labeled, links to the
pertinent info

- Have it all on a website

- Email to the subscriber list that this is available

- Update as more data come in
-----------------------------------------

[Brian] : In this use-case, there is a requirement to be able to search
        for, and aggregate, information across different archives that have
        data within the GRB error box. Each archive may have differing means
        of holding the information, (as mentioned "image" or "catalog") and
        we want to cross-correlate any "sources" that match our search.

        This use-case, taken strictly, at a minimum, this means that there must
        be well-described 'catalog' and 'image' structures out of which the
        search may pull 'data'. In order to do the cross-match scientifically,
        (by a central service, or the user) we will need to insure that in matching
        data the units are kept correct, that errors are propagated correctly,
        and that we can read this data. It would be desirable that the 'catalog'
        and 'image' formats share the same standard for describing this basic data
        (e.g. units/accuracy/type) in order to facilitate the cross-match
        fusion.



-----------------------------------------
2. Looking for new types of variable sources

- Identify the surveys/archives with multiple optical observations over
wide
areas (e.g., DSS, DPOSS, NEAT, SDSS....) in commensurate band-passes
(e.g., R
or r...)

- Find sources with a significant variability (defined in some
statistically
well defined fashion, e.g., more than 5 sigma and more than 1 mag
peak-to-peak amplitude, etc.)

- Cross-correlate these with other wavelengths and catalogs of known types
of sources (e.g., AGN, CVs, etc.)

- Develop supervised classifier for the variable sources on the basis of
the
known types (e.g., how to distinguish a QSO from a CV from a SN ...)

- Classify all variables accordingly, and identify those which do not fit
into any of the established classes

- Search on-line archives and surveys to find all available measurements
for these sources, characterize their spectral energy distributions, etc.
-----------------------------------------


[Brian]: This is another cross-archive, cross-match use-case. In this one, the
        user is trying to compare catalogs which have properties of variable
        sources.

        One requirement appears to be that the search must be executed
        over a variety of catalogs for some broadband magnitude. But are we
        to exclude catalogs of energy fluxes or spectra (photon or energy) from
        our search? Where does the conversion mechanism come in? Shouldn't it
        be done on the server/archive side where it might be more efficiently
        performed?

        Say, for example, our search is a V-band oriented one. Archives which
        have spectra for a given source that cover that bandpass should match, no?
        IF we create quantities with dependency on other quantities, this is
        possible. Consider that a magnitude and flux quantities for the visual
        bandpass could be defined as

                V = Fv = F(lambda)avg where lambda is range between 300 - 700 nm

        Now, as search for V will also return "Fv" as well as "F(300-700 lambda)"
        averaged spectral fluxes.

        Another point: There appears to be a requirement on the search that
        matching sources need to have returned information as the user expects,
        "find all available measurements for these sources". As a measurement may
        be a standard, but complex object, it would appear to be some type
        of thing that you can put in your query. This sort of thing may be done
        fairly trivially if the user may construct a structure, based on a standard
        quantity, to be "filled in" by the search.


-- 

  * Dr. Brian Thomas 

  * Code 630.1 
  * Goddard Space Flight Center NASA

  *   fax: (301) 286-1775
  * phone: (301) 286-6128