Quoting roles in data services

[Anne Catherine Raugh]

>From the PDS point of view, there is no "standard practice", really. We can
try to set standards, and for some specific roles we've done that, but we
have to fight the DataCite schema even to get the information in there, and
I don't think ADS is doing anything with it yet, although I'd be happy to
be wrong about that.

In the case of PDS the answer is straight forward, for now, because we
merely publish and distribute data. Once you put a service layer over it,
you add a layer of ambiguity as well. Citing the service is different from
citing the data, especially if the service does things like search and
collate products from multiple sources or produce plots, graphs, and
significantly enhanced images. More and more people who create services
would like to have their services cited or acknowledged. It's not clear to
me that "citation" is the right sort of relationship to ascribe to a data
service in the context of a research article about the data retrieved. For
example, when you finally find a copy of an obscure book from the LIbrary
of Congress with the help of a Special Collections Librarian, for example,
you may *acknowledge* the Library of Congress generally, or the Librarian
specifically, but you don't cite either one - you cite the book itself. (Or
more specifically, you cite the passages in the book that directly support
and reinforce the truth of your analysis.) The IVOA general case seems to
me to be similar to that, but the service creators and operators may see it
differently. These seem like they would be very muddy waters to wade into.

There are also intellectual property rights that might be involved. Even if
the data are public domain, if they are "significantly improved" (a topic
for later discussion and a determination that is often made by a court to
settle a suit), the "improver" can claim copyright on the result, which
means there might be a restrictive license required for what a user might
consider to be public domain data.

I'm on my way to Tucson for most of the week - meeting with the PSI folks
to plan our proposal, if and when NASA publishes an RFP.

-Anne.


On Mon, Feb 3, 2025 at 7:20 AM Baptiste Cecconi via semantics <
semantics at ivoa.net> wrote:

> Hi Stéphane,
>
> indeed, this is a bit of a tricky issue, since accurately acknowledging
> attribution or authorship is a key aspect of science ethics.
>
> My first answer is to clearly identify what contribution is attributed to
> whom. I mean: the EPNcore or ObsCore table metadata should be very explicit
> on what it is, and how to has been produced. The title and description has
> to be very clear on the content and the scope.
>
> With your example:
> The EPNTAP service is a "Curated catalogue of planetary observations by
> HST". The EPNcore metadata has been produced by Erard, et al. The original
> datasets used to produce the metadata are provided by the NASA/MAST
> database (or any other that has been used).
>
> The metadata in the Registry concerns the catalog metadata, and should
> (must?) contain reference to the progenitor data if required, e.g., with
> relation like "is-metadata-for" linking to the DOI (or any identifier) of
> the original dataset (possibly included as a provenance statement)
>
> So to me:
> - the metadata of the EPNcore/Obscore table should relate to the metadata
> and not the original data,
> - the relation to the source data should be explicitly stated and
> presented, both on the service description, as well as in the registry
> metadata.
> - the EPNcore/Obscore table landing page should propose explicit citation
> policies ("how to cite this metadata", "how to cite the underlying
> datasets")
>
> Cheers
> Baptiste
>
>
>
>
> Le 2 févr. 2025 à 11:01, Stéphane Erard via semantics <semantics at ivoa.net>
> a écrit :
>
> Hello
>
>
> I'm reviving an old request from the Solar System community but more
> general in scope:
> What are the common practices concerning references to data services? How
> do we cite creators of content (ie: data) vs services (ie, very often: a
> metadata table)? Can there be 2 different roles? - this is a recurring
> request from our providers, and a sensitive one.
> Example: I'm maintaining an EPN-TAP service identifying planetary data
> from the HST catalog. Creating the metadata table is a significant work
> which deserves identification and involves a responsibility; but I can't
> pretend I've created the data. How do I call my role there?
>
> VESPA distinguishes 3 roles, following current practices: Creator,
> Contributor, Publisher
> But there is an obvious ambiguity with Creator that needs to be addressed.
>
> Cheers
> Stéphane
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.ivoa.net/pipermail/semantics/attachments/20250203/0942a0ba/attachment.htm>