search

NCEAS / eml

Ecological Metadata Language (EML)
https://eml.ecoinformatics.org/
GNU General Public License v2.0
41 stars 15 forks source link

modify <dataset> to accomodate sibling <project>s #361

Open srearl opened 4 years ago

srearl commented 4 years ago

EML allows only one <project> element. However, there are occasions where a dataset is supported by or related to multiple, independent projects. In some cases, such documentation may be facilitated with <relatedProject> but this is inadequate when the projects are in fact not related. With the implementation of Structured Funding Information in EML 2.2.0, is it feasible to now consider supporting multiple, sibling <project>s such that funding elements from multiple projects can be documented and indexed?

cc @mobb

mbjones commented 4 years ago

We've had discussions in the past about allowing repeated projects. It would certainly be useful. Seems fine to me to relax the cardinality from {0,1} to {0,n} (which would need to be done in the various resource types including dataset. This would require a new schema release, and as it is a feature addition, the appropriate version for this would be 2.3.0, or, if it can wait, then maybe 3.0.0. Let's discuss if there is someone who wants to put together this change and the release, and if there is widespread support for it.

srearl commented 4 years ago

Thanks @mbjones, happy to contribute to that discussion.

mbjones commented 4 years ago

Sounds good @srearl . I think one question is whether people are ok with a new release right on the heels of the 2.2.0 release. Comments from folks would be welcome on this proposed change and whether a new release would be ok soon, or whether we should wait for more change request to accumulate over the next year. Thoughts on timing anyone?

srearl commented 4 years ago

Thanks @mbjones. From an LTER IM perspective, I do not think there would be any concern about the small gap between releases but it does seem excessive for only this change. I can squeak by with relatedProject. I guess the questions then are if you have additional changes you or others would be interested to address at this time, and/or whether you anticipate new requests coming in over the next year or so. I am happy to explore addressing this now and could inquire with the IMC if you would like input, but do not mind putting my particular suggestion on the wish list for the next update.

mbjones commented 4 years ago

@srearl There is a longer list of identified issues for EML 3.0.0 here: https://github.com/NCEAS/eml/milestone/22 Please review. The two labeled as breaking change would need to be in a 3.0.0 release and would probably break a lot of software. Some of the others could probably be done in a 2.3.0 release with far less impact. And yes, feedback or contributions on these issues would be welcome. I don't have funding or cycles to work on another release at the moment, and I'm mostly happy with EML 2.2.0 which will take us a while to catch up with in our software support. But I would support others if they had an itch they wanted/needed to scratch.

charuleka commented 3 years ago

Late to this discussion, but ESS-DIVE does have a need to support multiple projects in the schema. DOE tends to fund short-term (1-3 year) university projects with the expectation that they will partner with the long-term national lab projects (this doesn't really fit under the definition of a related project, since a university project can collaborate with multiple lab projects). Besides the national lab projects are also expected to collaborate with and leverage each others' efforts, but they may be unrelated projects with different science objectives.

We have a few datasets currently in ESS-DIVE that are listed under one project, but involved efforts from other projects (currently the other projects are only mentioned in the description).

ymgan commented 2 years ago

Hello, very late to this discussion, the antarctic community has many datasets that need the support for multiple projects. We are publishing biodiversity datasets to GBIF and OBIS in Darwin Core Archive which contains an eml file. Currently we are listing other projects in the description. I am very pleased to see that this feature is being requested. Thank you very much!