How to handle Material Groups in RISM

[TTG3333]

In RISM, some sources have one or more material groups (e.g. https://rism.online/sources/1001129868). In the data dump, they are represented with a hasMaterialGroup triple to a material group blank node. The material group blank nodes have a rdfs:label triple indicating the name (always of the format "Group 01"), and multiple hasSummary triples, each pointing to a summary blank node. The summary blank nodes represent a property of the material group, and have an rdfs:label triple indicating the name of the property, and a rdf:value triple indicating the value of that property.

In the spirit of reducing the amount of blank nodes and dummy URIs, the current plan is to have a material group entity (of the type https://linkedmusic.ca/graphs/rism/MaterialGroup), and each summary blank node gets translated into a triple, with the rdfs:label of the summary being mapped to a property, and the rdf:value of the summary used as the object of the triple.

The main issue currently is that I can't find a good property to link sources to the material groups (or the other way around), and I'm also wondering if there isn't a better way to handle the material groups.

What are your thoughts on this @dchiller ?

[dchiller]

I just want to make sure I understand your suggestion first.

So we have a node with URI https://rism.online/sources/1001129868 for the source. We then have another node with a dummy URI https://linkedmusic.ca/graphs/rism/MaterialGroup/xyz representing the material group. This material group node serves as the subject of a bunch of triples that contain the "summary" relationships in the data. The thing holding you back is that there is no good property to connect the source node and the material group node (the one with the dummy URI). Is that right?

[TTG3333]

Correct, and I'm also wondering if there isn't a better way than using a https://linkedmusic.ca/graphs/rism/MaterialGroup/xyz dummy URI

[TTG3333]

Ich said skip this for now

[annamorphism]

How common is it for there to be multiple groups? The examples of multi-group things I found on RISM seemed like cases where somebody wanted to differentiate a few cases of "this book starts on parchment but the rest is paper" or "the first part is from 1840 and the second part is from after 1850", without necessarily clearly defining what the first or second "group" consisted of (and I think if I, personally, were representing it in WD, to have said "material: parchment [applies to: group 1,] material: paper [applies to: group 2]"). But maybe I like qualifiers too much.

[ahankinson]

I can add a URI if that's helpful?

[annamorphism]

I think that makes sense--it makes an official URI rather than a "dummy" ID to talk about, and it could be connected to the main source using "part of" (which seems to be what people are doing for manuscripts on Wikidata when this problem arises.)

[ahankinson]

I’ve added the URI, but it won’t be available publicly until later this fall (working on a big release ATM.)

The pattern for it is pretty straightforward through:

https://rism.online/sources/<source_id>/material-groups/<group_num>

<group_num> is always two-digit, with a leading 0 for numbers less than 10.

[ahankinson]

Multiple material groups are fairly common. They are fairly flexible ways of identifying different types of materials in a single source — it could be a different type of surface material, but it could also be different sizes, inserts, production methods (manuscript, print, etc.)

They are represented this way because of the underlying MARC. Since MARC puts these different statements in different fields, we need a way to say “these fields belong to group 2” vs “these fields belong to group 3”. The way this is done is tacking on the $8 subfield with the group number.

Here's a simple example from a source that has two material groups:

=300  ##$ascore$801
=300  ##$aparts$802
=500  ##$aunbekannte Stimmenzahl$802
=590  ##$ano further indication$b$802
=593  ##$aManuscript copy$bNotated music$801
=593  ##$aManuscript copy$bNotated music$802

You can see that there are two groups, $8 01 and $8 02. In this particular case, the difference is that group 1 is the score, and group 2 is the parts. They both have 300 and 593 fields, but group 2 also has a 500, and a 590 with additional information about the parts.

[annamorphism]

aha! thanks for the explanation @ahankinson!