Mulgara

To see this, take any ascii file, save it in emacs with latin-1 encoding, try to load it, get the error message.  Then save it as utf-8 and it loads fine.

Ideally one should be able to specify any input encoding, but at a minimum I would suggest support for any form of Unicode, the (16?) ISO Latin encodings, one or two of the standard Japanese encodings, maybe a Chinese and Russian (KOI-8?).

I'm not sure how one would specify this; it should probably be specified in an HTTP header.  I looked at the "SPARQL 2":http://www.w3.org/TR/sparql-features/#sparql-update new features draft and the "SPARUL":http://www.w3.org/Submission/2008/SUBM-SPARQL-Update-20080715/ stuff but I don't see any mention of charset stuff.