Even if not used the lang thing would have implied that we'd considered
the language issue rather than assumed the whole internet spoke
> -----Original Message-----
> From: Matthew J. Dovey
> Sent: 16 July 2004 12:13
> To: 'Z39.50 Next-Generation Initiative'; [log in to unmask]
> Subject: RE: diagnostic examples - revisited
> > > Date: Fri, 16 Jul 2004 09:26:03 +0100
> > > From: "Matthew J. Dovey" <[log in to unmask]>
> > >
> > > > Basically we do/will ignore the returned message text
> in the live
> > > > situation. Of course we have a user profile which tells
> > us directly
> > > > what language they want.
> > >
> > > Which was one of the arguments against putting the message
> > element in
> > > (we didn't have such in SRW 1.0).
> > Right. But this is another Dumb Client thing, right? Among the
> > assumptions that Dumb Clients make is that their users are
> English (or
> > at least can read English error messages!)
> Right - which is why I didn't think the lang attribute (or
> element or whatever) was a big deal.
> Servers could ignore it - apart from those who were really
> dumb client/multi-language friendly (probably Theo's); dumb
> clients could do a if lang=="whatever" display message type
> code fairly easily (even XSLT can do that!).