search

w3c / dxwg

Data Catalog Vocabulary (DCAT)
https://w3c.github.io/dxwg/dcat/
Other
144 stars 55 forks source link

Rights statements can be handled in multiple ways, which makes interop difficult #1444

Open agreiner opened 2 years ago

agreiner commented 2 years ago

I wonder if we could make it a little easier to handle licensing and copyright statements. This section currently seems to direct people in three different directions (dcterms, ODRS, or ODRL).

andrea-perego commented 2 years ago

The intention of the guidance section on Licence and rights statements is not to recommend the use of different approaches, but rather to indicate the approach to be used based on given requirements.

More precisely, the default / recommended approach is to use DCTERMS, whereas odrl:hasPolicy should be used only if the rights statement is specified by using an ODRL policy. Moreover, ODRS is indicated (in NOTE) as a way to extend (and not replace) the DCTERMS approach, in case additional information should be specified (e.g., attribution text, attribution URL).

@agreiner , do you think this should be made more explicit in that section? In such a case, do you have any suggestion?

agreiner commented 2 years ago

@andrea-perego , what you wrote above is a big help for understanding the intention. I think it would help to start the section with a very clear statement similar to that.

aisaac commented 2 years ago

Actually while we're at it, maybe we could present a sort of gradation, saying that ODRS is an alternative to DCTERMS, and ODRL goes one step further in the flexibility/complexity that it allows? For the current wording with ODRL is "in the particular case when rights are expressed via ODRL policies, it is recommended [...]". Maybe we could write "For more advanced cases, ODRL can be used to specify potentially complex policies. Then, it is recommended [...]".

makxdekkers commented 2 years ago

@aisaac I wonder if we should be in the business to give this type of advice at all. I think it is very hard for us to determine in which case any of the approaches are used. It is really up to a community to decide what they want to do. For example, I am not sure if ODRL is always only used for complex cases -- it could well be that a particular community uses ODRL for everything, including very simple cases, e.g. anything allowed/no restrictions. The current text just states that if ODRL policies are used, this is the way to do it.

agreiner commented 2 years ago

Hm, I keep thinking about this from the point of view of a developer. If all the data is using different ways to describe a very fundamental piece of metadata, it gets very difficult to handle it in an application. It's nice to be accommodating of the folks who are describing their data, but the value of a standard lies in having uniform inputs. Since including this type of information is crucial to data sharing, I'd really like to see us state a strong preference.

aisaac commented 2 years ago

Maybe there can be a more neutral way. Actually the current section makes it quite technical. First the various dcterms properties and what they're supposed to point to according to their standardized semantics, and then the ODRL policies which are just presented as "in the particular case when rights are expressed via ODRL policies." This is hard to argue with, no? Also I believe that it states quite well an order or preference. Simple statements first, and more complex policies second.

Maybe what makes @agreiner uneasy is the allusion to ODRS? It's already merely a note but I can see it stands in the middle of the story.

There I think we can turn to a rather neutral approach, or at least one that just follows from our W3C anchoring. ODRL is a W3C recommendation. So as a W3C group we should be expected to recommend existing W3C recommendations for specific cases. Which means that either ODRS (as a "more sophisticated approach to express rights") should be pushed towards the end of the subsection, or it should be removed.