Add “licence attribution” for source reference, licence number etc.

[effjot]

New field for license numbers, ODC-BY and German Data Licence attributions etc.

• New database text field dataset.license_attribution (I followed American spelling)
• Added field to Dataset attribute form below licence selection in Admin Project
• Added to HTML template

Implements #71

[mdouchin]

Thanks a lot for your contribution.

I wonder if we should add also the new fiels in the DCAT view. See: https://github.com/3liz/qgis-pgmetadata-plugin/blob/master/pg_metadata/install/sql/pgmetadata/30_VIEW.sql#L170

At present, this will return something like

<dct:license>Licence Ouverte Version 2.0</dct:license>

We should probably add the license URI in the XML, such as

    <dct:license rdf:resource="http://creativecommons.org/licenses/by-nc/3.0/"/> 

Could you also add a COMMENT on the new column, to explain what is expected ? See comments for other fields visible here: https://docs.3liz.org/qgis-pgmetadata-plugin/database/tables/dataset.html

See source: https://github.com/3liz/qgis-pgmetadata-plugin/blob/master/pg_metadata/install/sql/pgmetadata/70_COMMENT.sql#L158

[mdouchin]

After some thinking, I am not very fond of adding a new field in the dataset table: it increases complexity and does not bring much advantage compared to the existing situation.

I see two ways to address your need:

• Add the needed line(s) in the glossary with your new license(s). See for example what have been done for the French "License ouverte"

wkt_geom	id	field	code	label_en	description_en	item_order	label_fr	description_fr	label_it	description_it	label_es	description_es	label_de	description_de
NULL	128	dataset.license	LO-2.0	Licence Ouverte Version 2.0		0
NULL	129	dataset.license	LO-2.1	Licence Ouverte Version 2.1		0

We have one line per license and version. We could improve the views (DCAT, HTML, etc.) by adding the code close to the license, for example Licence Ouverte Version 2.1 (LO-2.1).

You could add these lines in your own database, or propose a PR to add them in the source code for all German users (the latter is my preferred solution of course)

• Move the license items from the glossary to a new dedicated table. This would help to add more information about licenses, such as URL, date of validity, etc. The views could concatenate this fields and create a more detailed license text. It seems a little bit overkilled, and we will need to warn the user to update their QGIS administration project.

[effjot]

I’m not sure adding to the glossary would be a good solution, because it would fill the glossary table with a lot of custom entries and make upgrading a nightmare.

My use case is this: for any given license, there might be an attribution which we need to show on published maps, and we have to check if we could use the data in a given project. Data could be licensed under ODC-BY or the German Data Licence, however, the “by” part of those licences requires the attribution to be shown, and could be any text (although some attributions would appear more frequently). For example:

• Landesamt für Umwelt Brandenburg → just the name of an agency; I expect about 10–15 of those to appear frequently in our daily work
• Geobasis-DE/LGB 2020 → another agency attribution but with year
• Genehmigung GB-G 06/19 → license number (usually data that has been provided for a concrete project only)
• Mail LfU Hans Meier 01.01.2021 → source for project-only data

So my suggested database field would be some ugly mixed bag. Moving the whole license stuff to a separate table might be cleaner, but bigger effort and change for the users, as @mdouchin remarked.

I don’t have a deeper knowledege of metadata standards. Maybe this problem has already been adressed somewhere?

[mdouchin]

Thanks for your examples which help to understand.

The real questions is : do you need to adapt the license for each dataset, meaning the same license can have different text attribution for different datasets, or are the complementary information only depends on the license (and version) ? It will determine if we need to add a new field in the dataset table or not.

I think the new "License" table won't be too hard to implement, as it will be with a simple foreign key.

[effjot]

The real questions is : do you need to adapt the license for each dataset, meaning the same license can have different text attribution for different datasets, or are the complementary information only depends on the license (and version) ? It will determine if we need to add a new field in the dataset table or not.

It is the first case. Not many different licenses (license types, e.g. “Licence Ouverte Version 2.1” or “Data licence Germany – attribution – version 2.0”), but for any license many different attributions (for CC-by licenses) / sources / license numbers (for governmental/commercial data) can exist, depending on the dataset. So I think the licenses (license types) definitions could be kept in the glossary or a new “License” table. But the attributions belong in the dataset table, in my opinion.