- From: Cristiano Longo <cristianolongo@opendatahacklab.org>
- Date: Fri, 23 May 2025 10:54:26 +0200
- To: public-ontolex@w3.org
Yes, you are totally right. An example can be found in the ontolex lemmon specification (https://www.w3.org/2016/05/ontolex/#written-representation-datatype-property) about the English word color, which have two written representations (color and colour) with two different language tags (en-US and en-GB, respectively). Thanks all, CL On 23/05/25 10:37, Gilles Sérasset wrote: > Hi all, > >> I must say that, may be just in my case, sounds quite redundant indicating the language both in the lexicon and in all the written representations. > I also thought it was in initial versions of DBnary, however, I witnessed people querying on langStrings and filtering on the language. > > I get that it allows for direct checking of the correct language without link resolution to go up the graph to the lexicon/entry language (and many endpoint may even index the lang to avoid an extensive a posteriori language filter). > > But the real reason is also that a lexicon may (and frequently do) contain information on several languages/dialects in the same form. > > E.g. in Chinese DBnary all Chinese entries are associated to many phoneticRep in different Chinese languages/dialects. The same is also possible in any other language/dialect. > > Anyway, the problem you mentioned has been discussed in several contributions (e.g. for some South African languages). > > Regards, > > Gilles,
Received on Friday, 23 May 2025 08:54:41 UTC