Mapping to W3C and ISO standards

[antaldaniel]

I have been looking at frictionless for a long time, and I must admit that I have so far not understood the advantages of chaning from a standardised way of describing and releaseing a dataset to a non-recognised standard. I have tried to understand the documentation, but apart from "inspiration" from DCAT and DataCite, it is not clear to me if you aim to be compatible with metadata standards?

I am developing an R package (called dataset) that aims to keep the dataset's abstract strucutre (a data.frame), its serialisation and all metadata in one object, as opposed to creating a package of datasets about data and datasets about the dataset. One rOpenSci reviewer asked to explain the difference of this approach to frictionless, and of course, early-binding is the difference, but I could see value in creating frictionless transition to frictionless, if I understood more the vocabulary of the properties you use.

While my package is strictly R only, it aims to allow the user to serialise datasets with full metadata annotation following several standards, including DCTERMS, DCAT (and the EU DCAT-AP) and also DataCite, and to provide serialisation for publishing on the EU Open Data Portal or Europeana.

[roll]

Hi @antaldaniel!

Thanks for the question. I'm cc @peterdesmet here as well

[antaldaniel]

@roll , super, and I would really be happy to have a fruitful discussion, because I could probably also help to implement things. An rOpenSci reviewer brought to my attention that our goals are very similar, but we come from a strictly information science / statistics background and we started to develop something similar to you, with a very strong emphasis on the expectations of the EU Open Data Portal, statistical exchanges, libraries and open science repositories (i.e., SDMX, DataCube, DataCite, DCTERMS, DCAT.) It is a big overlap with your "inspirations", and we took it more like strict requirements.

Where our ambition is far smaller is that we only work in R and RDF; where perhaps a bit greater is that we also include in a similar data package to yours multimodal assets (photographs, scanned documents, music files), in these cases relying for example on the new RiC standard that replaced six former ISO archival standards in 2023.

I could see many ways of alignment, partly perhaps making your requirements slightly more strict, and partly by developing recommended option that are more strictly following the standards governing long-term data storage places, i.e., libraries, archives, museums and their digital repositories.

[peterdesmet]

Hi @antaldaniel, sounds interesting! My background is also in science. It's probably easier to discuss this in an online call, can you reach out to me via email? You'll find my email address at https://orcid.org/0000-0002-8442-8025