Re: Localization of RDF property names? Seeking advice.

Peter and Richard,

It is my (possibly flawed) understanding that RDF can use IRI's
(Internationalized Resource Identifiers) [1] in place of URI's.  This
practice would allow the use of most Unicode characters directly in your
URI's.

I'm not sure how well this works in practice, though.  There are likely
to be problems with much of the existing Semantic Web software base.  I
would expect that most existing software packages will "almost" work
when using UTF-8 encoded IRI's in place of URI's, but that they will
suffer from a number of bugs wherever they do things like URI parsing or
validation.

Cheers,

Ian

[1] http://www.ietf.org/rfc/rfc3987.txt

-------- Original Message  --------
From: Richard Cyganiak <richard@cyganiak.de>
To: "Peter Krantz" <peter.krantz@gmail.com>
Cc: semantic-web@w3.org
Subject: Re:Localization of RDF property names? Seeking advice.
Date: 12/17/2006 2:33 PM

> 
> Peter,
> 
> On 17 Dec 2006, at 13:40, Peter Krantz wrote:
>> We are using RDF to make statements about documents in an information
>> domain where it is difficult to translate concepts to english (some do
>> not have a corresponding english word). Should we try to translate
>> these concepts to english or leave them as they are (and remove
>> accented characters)?
> 
> There are two distinct issues here. First, what URIs to use for your
> concepts? Second, what rdfs:label to use for the concepts?
> 
> The second one is the smaller problem, you can use literals with
> language tag to attach multilingual labels to concepts, and all Unicode
> characters are allowed. So, provide labels in all languages that users
> of your application are likely to need.
> 
> The first question, what URIs to use, is harder. You should keep in mind
> though that end users should normally not see the URIs, that's what the
> labels are for. So, in theory you could avoid the problem and use URIs
> like http://my-namespace/123 instead of
> http://my-namespace/difficult-to-translate-concept. But that would be
> hard on the people who work with the URIs directly -- ontology
> engineers, software developers, and the like. Using mnemonic URIs is of
> course much more convenient for them.
> 
> So I guess you should ask yourself if everyone who is likely to work
> directly with the RDF is able to understand the untranslated concept
> names. If the answer is yes, then I think using the original language in
> the URIs is a good solution.
> 
> Richard
> 
> 
>>
>> Is there anyone else that has had the same question and what was your
>> approach?
>>
>> Kind regards,
>>
>> Peter