- From: Gil Francopoulo <gil.francopoulo@wanadoo.fr>
- Date: Fri, 6 Nov 2015 10:07:34 +0100
- To: public-ontolex@w3.org
dear Laurent, Philippe and Felix, I thought that we already decided to follow IETF BCP47, let's recall the talk with Felix and Philip on this list (it was last year, no ?). I'd like to add another point to Laurent's message: IETF BCP 47 is a little bit more than 2 letters when they exist, 3 letters otherwise. This is definitly true, but BCP 47 is a set of rules on how to combine ISO lists of codes. For instance, when you want to express a language mark for English as used in the USA, you can combine a language code (always in lower-case, according to ISO) with a country code (always in upper-case, according to ISO) like this: en-US. To finish, if you just refer to ISO-639-1 or ISO-639-3, how do you do? You cannot express this. Bonne journée, Gil Le 05/11/2015 10:58, Laurent Romary a écrit : > Dear all, > Would it make sense to follow the W3C recommendation for xml:lang, which is http://www.rfc-editor.org/rfc/bcp/bcp47.txt ? > (2 letters when they exist, 3 letters otherwise) > Bonne journée à tous, > Laurent > > >> Le 5 nov. 2015 à 10:49, Gerard de Melo <gdemelo@mpi-inf.mpg.de> a écrit : >> >> Hi Philipp, >> >> For Lexvo.org, you can either use 3-letter ISO 639-3 codes such as >> >> http://lexvo.org/id/iso639-3/eng >> >> or 2-letter ISO 639-1 codes such as >> >> http://lexvo.org/id/iso639-1/en >> >> Gerard >> >> >> On 2015-11-5 17:03, Philipp Cimiano wrote: >>> And one small thing: >>> >>> in many examples we use the dublin core terms property "language" to >>> express language information. >>> >>> Sometimes we use "eng" and "en" as language identifier (local name) with >>> respect to the LOC and LEXVO datasets. >>> >>> It seems to me that there are some mistakes here. >>> >>> It seems that LOC uses "eng" and LEXVO uses "en". >>> >>> For instance: >>> >>> http://www.lexvo.org/page/iso639-1/eng -> does not resolve property, while >>> >>> http://www.lexvo.org/page/iso639-1/en -> does resolve >>> >>> On the other hand: >>> >>> http://id.loc.gov/vocabulary/iso639-2/en -> does not resolve, while >>> >>> http://id.loc.gov/vocabulary/iso639-2/eng -> does resolve >>> >>> Thus, in all examples we should use: >>> >>> lexvo:en and loc:eng to indicate language information. >>> >>> A minor thing, but still... >>> >>> Greetings, >>> >>> Philipp. >>> >>> >>> >>> Am 05.11.15 um 09:38 schrieb Philipp Cimiano: >>>> Dear all, >>>> >>>> apologies for my silence in the last weeks. I have been distracted by >>>> many other things in October. Now I am working on the release of the >>>> ontolex spec with high priority. I have worked through all the >>>> normative sections of the specification. There are smaller issues that >>>> I will clarify with the main stakeholders of these modules directly. >>>> >>>> It is now time to have a final check of the spec if you want. >>>> >>>> I am confident that I will finish the thorough check of the model in >>>> the next few weeks. >>>> >>>> I will then give all the opportunity of doing a final check in the >>>> first two weeks of December. >>>> >>>> We will then release the specification of the model this year. >>>> >>>> Kind regards, >>>> >>>> Philipp. >>>> >>>> > Laurent Romary > INRIA > laurent.romary@inria.fr > > > > > > >
Received on Friday, 6 November 2015 09:08:33 UTC