Re: New attempt

----- Original Message -----
From: "Benno Blumenthal" <address@hidden>
To: "John Caron" <address@hidden>
Cc: <address@hidden>
Sent: Wednesday, June 05, 2002 1:32 PM
Subject: Re: New attempt


> John Caron wrote:
>
> > One thing I dont understand is what is the relationship between whats
> > contained inside the collection vs. whats in the dataset you specify
through
> > the access element? The user should understand them to be the same
(although
> > in fact the writer could accidentally or on purpose make them
unrelated)?
>
> It is not a dataset that I specify through the access element -- it is a
> DODS access for the collection (collections are the same as datasets -- we
> cannot even formulate a reasonable distinction).  Most collections would
not
> have DODS access, but some do.  The writer cannot make them different,
that
> violates the meaning of access belonging to the collection.

Sorry, I was using the word dataset rather sloppily to indicate both the
THREDDS dataset element, and the "abstract thingy referred to by a DODS
URL".

To rephrase my question:

What is the relationship between whats contained inside the THREDDS
collection element vs. what you get back from the abstract thingy?

>
>
> > I
> > presume that the ANNUAL, MONTHLY, and SEASONAL datasets are a different
way
> > of accessing the collection dataset? Or should they be thought of as
> > different, but related datasets?
>
> They are subdatasets, i.e. different ways of accessing the collection
> dataset.

It sounds like they are different datasets (abstract thingy), in the sense
that you would get different data then you would from the original
collection data.

> In practice this is important because some DODS clients can handle
> DODS datasets that contain nest variables (i.e. structures), and some DODS
> clients can only handle flat datasets, i.e. a collection of simple
variables.
> The DODS clients that can handle nesting can use the collection access
through
> DODS:  DODS clients that cannot need to drill farther before using DODS.

Its not obvious how a DODS client would know it needs to drill down. There
is a valid DODS URL at each level.

>
> > In this spscific example, you also specify that the collection has a
service
> > of type Catalog <service serviceType="Catalog" suffix="thredds.xml"/>.
What
> > is the relationship between that catalog and what is contained in the
> > collection?
>
> a catalog with a single collection is the collection:  your formulation as
best
> I can understand it (i.e. I see no different between catalogs that contain
> single collections and collections, which is implicit in your dropping the
> level of nesting under that circumstance).  I think I should omit that
circular
> reference (not that the client would get too confused:  when it
constructed the
> url it would find it the url of the document),
> e.g. the file should be
>
> <?xml version="1.0" encoding="UTF-8"?>
> <!DOCTYPE catalog SYSTEM
> "http://www.unidata.ucar.edu/projects/THREDDS/xml/InvCatalog.0.6a.dtd";>
> <catalog>
> <service ID="IngridDataset" serviceType="Compound"
> base="http://iridl.ldeo.columbia.edu/";>
> <service serviceType="DODS" suffix="dods"/>
> <service serviceType="Catalog" suffix="thredds.xml"/>
> </service>
> <service ID="JustDODS" serviceType="DODS" suffix="dods"/>
> <dataset name="LEVITUS94">
> <attribute name="fullname" value="LEVITUS94"/>
> <access serviceID="JustDODS" urlPATH="SOURCES/.LEVITUS94/"/>
> <attribute name="references" value="Conkright:etal1994 Levitus:Boyer1994a
> Levitus:etal1994 Levitus:Boyer1994c"/>
>
> <documentation>LEVITUS94:  World Ocean Atlas 1994, an atlas of objectively
> analyzed fields of major ocean parameters</documentation>
>
> <documentation
>
xlink:href="http://iridl.ldeo.columbia.edu/SOURCES/.LEVITUS94/.dataset_docum
entation.html"/>
>
> <document
>
xlink:href="http://iridl.ldeo.columbia.edu/SOURCES/.LEVITUS94/.oceanviews.ht
ml"
> name="oceanviews"/>
> <document
>
xlink:href="http://iridl.ldeo.columbia.edu/SOURCES/.LEVITUS94/.oceanviews2.h
tml"
> name="oceanviews2"/>
> <document
> xlink:href="http://iridl.ldeo.columbia.edu/SOURCES/.LEVITUS94/.Zmix.html";
> name="Zmix"/>
> <dataset name="ANNUAL">
> <access serviceID="IngridDataset" urlPATH="SOURCES/.LEVITUS94/.ANNUAL/"/>
> </dataset>
> <dataset name="MONTHLY">
> <access serviceID="IngridDataset" urlPATH="SOURCES/.LEVITUS94/.MONTHLY/"/>
> </dataset>
> <dataset name="SEASONAL">
> <access serviceID="IngridDataset"
urlPATH="SOURCES/.LEVITUS94/.SEASONAL/"/>
> </dataset>
> </dataset>
> </catalog>
>
>
> Benno
>
>
> --
> Dr. M. Benno Blumenthal          address@hidden
> International Research Institute for climate prediction
> Lamont-Doherty Earth Observatory of Columbia University
> Palisades NY 10964-8000                  (845) 680-4450
>
>
>