RE: native encoding

No.  If I ask for a UKMARC record, I shouldn't have to specify a variant as
well.  The UKMARC specification defines its characterset.  Nothing we say in
a Z39.50 request can change that.

We are going to have to profile (through an implementors agreement) which
record syntaxes the UTF-8 negotiation applies to.  Personally, I do not want
the UTF-8 negotiation to apply to USMARC records, even if it is
theoretically possible to get them UTF-8 encoded.  I expect to get them
encoded in USM-94 (ANSEL+EACC).

Record Syntaxes that should be effected by UTF-8 negotiation:
SUTRS

GRS-1 could be on that list, but it is closely tied with the variant
specification and I don't feel badly about using that to control the GRS-1
characterset.  (But, I have never supported GRS-1, so do whatever you want
with it.)

Ralph

> -----Original Message-----
> From: Ray Denenberg [mailto:rden@loc.gov]
> Sent: Friday, March 01, 2002 10:12 AM
> To: zig
> Subject: native encoding
> 
> 
> The character encoding discussion seems now to
> focus (and I use that term loosely) on native
> encodings, that is, if we negotiate utf-8 for a
> session and if a particular syntax has a
> well-known, native encoding other than utf-8,
> which applies?
> 
> Perhaps I missed something and if so please
> refresh my memory:  What is the objection to using
> variants?
> 
> Thus if utf-8 is negotiated it applies to
> everything unless explicitly overiden. If you want
> to request a record in an encoding other than
> utf-8, you include a variant request; if a server
> wants to supply a record in an encoding other than
> utf-8, it includes a supplied variant.
> 
> Please, if anyone objects to this approach speak
> up.
> 
> --Ray
> 
>