Dear all, There was a homework about the mapping section of the CRM Site which is out of date and needs reworking and rethinking. Everyone is quite busy right now and there are quite a few things to consider in that work anyhow.
SO as a homework, I present to you the conversation we have opened. Here it is pasted as a series of interjections by myself Eleni and Pavlos. In an issue management system, this would be easier to follow (see related homework). Quoth George: "Dear Eleni, all, This section presents quite a challenge. The material it contains is uneven—mixing different positions (mostly FORTH’s), examples of mappings, strategies for mappings, PowerPoint sketches, and general ideas. A key difficulty is that the notion of *“mapping”* here is not well defined: - CIDOC CRM itself is not a fixed target model. Its constructs are too open-ended, meaning that mappings inevitably vary project by project and dataset by dataset. - In some cases, the section seems to suggest “model-to-model” mappings (e.g., LIDO 1.0 → CRM), but in reality, no project implements a source model perfectly, so this path is more theoretical than practical. - At the technical level, standards like *X3ML* *do* express mappings. But it is unclear whether this section intends to cover such technical mappings, or the looser model-to-model alignments. If the goal is to present model-to-model mappings, we should consider: - Defining what we mean by an “ideal” mapping. - Suggesting acceptable formats (X3ML or equivalent). - Making clear that multiple technical solutions are possible. If instead the goal is to discuss the practice of modelling, this belongs in the teaching section, which would require new text (and, realistically, time and authorship). As it stands, the section is outdated and inconsistent. A better approach might be to frame examples around concrete cases where mappings *to ontology profiles* are required—for instance: - *Sealit*, which had defined models as transformation targets. - *Linked.Art*, which provides examples like this CDWA mapping: https://linked.art/cookbook/mappings/cdwa/ (though still somewhat loose in format). My concern is that clarifying and reworking this section could be quite a large project. Still, it seems necessary if we want it to be useful and consistent. What do you think—do you see this as worth pursuing? Best regards, George" Quoth Eleni: "Dear George, Admittedly, I had not given it much thought until now. I see what you mean with the different perspectives. Maybe before we embark on another 3year project, we run this by the SIG saying we could split the Mappings section in 3 parts (your suggestion actually) 1. Defining what we mean by an “ideal” mapping. 2. Suggesting acceptable formats (X3ML or equivalent). 3. Making clear that multiple technical solutions are possible and use the existing mappings as blueprints for model-to-model, technical mappings, etc. but for later versions. Not knowing much about the "data examples" that were xml files mapped to some version of CIDOC CRM (currently under Resources\Data examples <https://cidoc-crm.org/data_examples>), we probably need to revise those too. Cause I don't know how helpful they are for CRM users. The Synergy model needs to be updated too probably... So let's start small, get the SIG's opinion, and give ourselves the time to go over things properly, rather than come up with an ad hoc solution that still lacks perspective. Cheers Eleni " Quoth Pavlos: "Hi all, We could probably work on this after the restructuring of the website has been completed and agreed. But for sure, it needs an update. The content is very very old. I think we should first distinguish between model-to-model mappings and data transformation mappings, and then, for each one, provide some theory, tools, etc. Best, Pavlos" So as it is plain to see we are not at a decidable proposal stage, but I think it is interesting to open up this issue to a broader discussion group here on the list, if you have ideas, or at SIG if there is time, to address some of the theoretical questions that we might want to agree on before moving forward more concretely. Thanks and look forward to hearing your ideas! Best, George
_______________________________________________ Crm-sig mailing list Crm-sig@ics.forth.gr http://cidoc-crm.org/crm-sig-mailing-list
