You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When authoring shacl, the subject ordering of turtle-formatter gets in the way when named property shapes are used. Usually, property shapes are written as blank nodes, but if they occur multiple times, it is often better to give them their own URIs and make them 'top-level entities', so to speak. Now for ease of editing you would like to keep those property shapes close to the node shapes that they are used in. Otherwise, when making one change, a lot of scrolling or searching in the file may be required, which increases the cognitive load on the person doing it (ie, it's a pain in the neck) and may lead to editing errors.
Similar requirements may exist in other modelling domains, where you might want to keep e.g. members of a family close togehter, or maybe parts of an engine in close proximity of each other.
The current ordering options do not support such use cases very well.
I might be interested in providing a solution, but if so, it will take time. Meanwhile, if anyone has additional requirements or ideas for an algorithm (@ralphtq) , please comment!
The text was updated successfully, but these errors were encountered:
Hi @fkleedorfer,
I understand the request. However, I'm not convinced yet that this is SHACL-specific. Can we generalized the topic to: "Add an option to keep existing subject order", i.e., not have the subjects sorted?
What is needed for that is the ability to keep line information (provided by a custom parser) with each resource. After that, it's quite trivial; you just render resources sorted by their original line. I've commented about this here on issue #2 (because retaining comments also requires custom parsing). As mentioned in that comment, incidentally I've already implemented this approach in the Eclipse Semantic Modeling Toolkit so I know it works :-) but the implementation there is not generically applicable to RDF.
When authoring shacl, the subject ordering of turtle-formatter gets in the way when named property shapes are used. Usually, property shapes are written as blank nodes, but if they occur multiple times, it is often better to give them their own URIs and make them 'top-level entities', so to speak. Now for ease of editing you would like to keep those property shapes close to the node shapes that they are used in. Otherwise, when making one change, a lot of scrolling or searching in the file may be required, which increases the cognitive load on the person doing it (ie, it's a pain in the neck) and may lead to editing errors.
Similar requirements may exist in other modelling domains, where you might want to keep e.g. members of a family close togehter, or maybe parts of an engine in close proximity of each other.
The current ordering options do not support such use cases very well.
I might be interested in providing a solution, but if so, it will take time. Meanwhile, if anyone has additional requirements or ideas for an algorithm (@ralphtq) , please comment!
The text was updated successfully, but these errors were encountered: