[OSGeo-Standards] DOIs for Datasets

[Rob Atkinson]

DOIs only really need the resolver to be hosted - so a DOI should resolve
to a local resource identifier (IRI) - which itself should be permanent
IMHO - and resolve to a URL for the current location of the resource.

This last redirect is important to allow different flavours of the resource
to be served with content negotiation.

And of course we have the resource descriptor (a DOI landing page) as well
as the data itself.

This is best (i.e. if we stick to standards that meet long term needs)
handled using the DCAT model of a Dataset and a set of associated
DataDistributions - but legacy style fixed and limited functionality schema
based solutions (e.g. XML schemas) could co-exist for specific content
types - it just wouldn't a great investment today to implement
infrastructures based entirely on specific technologies (physical model)
rather than a technology neutral conceptual framework.  (SDI 1.0 has
suffered greatly from this practice, hence the need for this upgrade..)






On Mon, 26 Aug 2024, 06:41 Joana Simoes via Standards, <
standards at lists.osgeo.org> wrote:

> Good evening,
>
> I was wondering if anyone has experience/best practices to share
> regarding generating DOIs for datasets that are hosted on a SDI?
>
> Research resources, including data, are encouraged to adopt Digital
> Object Identifiers (DOI) to identify objects persistently. They allow
> things to be uniquely identified and accessed reliably.
>
>  From what I understand, these DOIs need to be generated by one of the
> registration agencies under the DOI foundation (https://www.doi.org/).
> Some of these agencies, like Zenodo (https://zenodo.org/), will generate
> the DOIs for free, but the issue I see is that they require the data to
> be hosted on their repositories. I see this as a duplication of
> information if the data already lives on a SDI, raising a series of
> issues (maintaining the data in both places, duplication of storage, etc).
>
> What would be the best approach, to keep the data (and metadata) on the
> SDI, but still be able to reference it as a citation?
>
> In this context, would it make sense to give up on the DOI and go for a
> Persistent Uniform Resource Locator (PURL)?
>
> This would be an example of a Purl:
>
> https://doi.byteroad.net/emotional.cities/hex350_grid_cardio_1920
>
> Looking forwards for your thoughts on this.
>
>                                    Thank You,
>
>                                         All the Best,
>
>                                                  Joana
>
> _______________________________________________
> Standards mailing list
> Standards at lists.osgeo.org
> https://lists.osgeo.org/mailman/listinfo/standards
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/standards/attachments/20240826/9065bb98/attachment.htm>