search

w3c / rdf-dir-literal

Proposal to add base direction to RDF Literals
Other
8 stars 6 forks source link

A point against `LocalizableString`: redundancy #2

Open pchampin opened 5 years ago

pchampin commented 5 years ago

While my first idea was indeed to introduce a new datatype (as the least invasive solution). But I came to think that it would be a very bad solution. It would recreate the "plain litteral / xsd:string" dilemma that plagued RDF 1.0 . We will have a mix of LocalizedStrings and langStrings, creating a mess for data consumers.

By avoiding to break standards and implementations, we may break the data ecosystem, which is much worse.

iherman commented 5 years ago

True. But it is not impossible that, on long term, an RDF WG working on RDF 2.0 could make this redundancy disappear, just as RDF 1.1 equates plain literal and xsd:string. So LocalizableString could be seen as a temporary solution as far as the core RDF model is concerned, but it would allow to deploy updated serialization formats, including SPARQ syntax. Ie, for most of the users the issue would become transparent...

pchampin commented 5 years ago

Equating plain literals with xsd:string worked without changing the concrete syntaxes. That's what made the transition so smooth -- even unnoticeable, for some.

LocalizableString would introduce a new way to write literals, distinct from langString. This will be much harder to mend, IMHO.