I vote this. While I was implementing IPP, I felt we have too many
data types for text attributes. We have charcter-set, natural-language,
name and text with/witout natural language and keywords...it's too enough.
And I think the natural language override mechanism is very confusing...
it could be a potensial interoperability problem. Although "always use
texts with language" could cause other problems, I belive it would solve
more problems than it will cause.
--------
Yuji Sasaki
Company E-Mail :sasaki at jci.co.jp
Personal E-Mail:crazy17 at ibm.net
Nifty-Serve :PFG02524 at niftyserve.or.jp