Re: Small items for SPARQL

Dan Connolly wrote:
> On Fri, 2004-12-17 at 21:21 +0000, Seaborne, Andy wrote:
> [...]
> 
>>I will follow this trend and put in rq23/ that the grammar is over the 
>>UNICODE character set,
> 
> 
> yes, very well, but...
> 
> 
>> and say that the files are encoded in UTF-8.
> 
> 
> Hmm... what files are discussed in rq23/?

Discused?  None directly but it gives the grammar. "the files" was referring 
to "files on disk" from the start of the text.  I don't care where the 
statement goes but it should go somewhere - next to the grammar is as good a 
place as any as far as I can see.  It is not just a protocol issue, nor just 
a test files issue.

It does not work to just not say anything - one app runs with ISO-8859-1, 
one on UTF-8 and it does not work e.g character á.  At best, the ISO 8859-1 
char becomes an encoding error in UTF-8.

	Andy

PS The convention for queries has been to use file extension .rq

> 
> It makes sense to document the contents of the test
> files in the test suite description, and it makes
> sense to specify encodings in URIs and bodies in
> the protocol...
> 
> are you going to add an Internet Media Type section?
> I suppose that would make sense.