*** posted on behalf of Harry Wagner ***

Dear Shigeo,
Thank you for the information, and for your previous note.  I am on vacation
until September 4, but will look at the information you provided over the
next couple of days.  The Registry prototype
(http://wip.dublincore.org:8080/registry/Registry) is also using unicode
(UTF8) for it's encoding scheme.  In order to differentiate between
different language versions of words like "titel" we are using different
databases.  The UI provides an input field for selecting a language.  That
field is what tells the prototype what database to search.

I hope to have Japanese loaded in the next week or two.  The German and
Chinese that is loaded is not the correct text, but is only there to prove
that we can display and search in multiple languages.  I would appreciate
hearing any comments you have regarding our approach to the multilingual
requirement.  Thanks again.  This exchange has been very helpful.

Best Regards,
Harry Wagner

>
> ----- Original Message -----
> From: "Sugimoto Shigeo" <[log in to unmask]>
> To: <[log in to unmask]>
> Cc: <[log in to unmask]>; <[log in to unmask]>
> Sent: Monday, August 27, 2001 9:36 PM
> Subject: Re: Prototype Changes
>
>
> > Dear Harry,
> >
> > My response is a bit late...
> >
> > On Thu, 23 Aug 2001 08:32:23 -0400, "Wagner,Harry" <[log in to unmask]>
> wrote:
> > > ULIS (http://www.dl.ulis.ac.jp/DCML/DCML.html) appears to have taken
> three
> > > different approaches, based on the level of browser support.  If the
> browser
> > > supports the language font they produce the appropriate unicode (this
is
> > > what the registry prototype does now).  If the browser does not
support
> the
> > > language, but supports Java then they produce the translation using a
> Java
> > > applet.  If neither of these apply they display a graphic (.gif or
> .jpg).
> > >
> > > My view is that the browser support for languages has matured enough
> that we
> > > can expect our users to be using a browser that supports their
language.
> > > Both Netscape and IE provide this function, and are both free.  I am
> hoping
> > > we can discuss this during the conference call this morning.
> >
> > As I mentioned in my previous message to you and CCed to DC-REGISTRY,
> > the ULIS registry has two versions. The URL referred above is our very
> > first version. In the second version, we are using Unicode as its
> > underlying character encoding scheme.
> >
> > In the second version, we provide a simple search function using Uniode
> > and XML/XSL.
> > This function is based on simple text matching on Unicode.
> > Global IME on Windows is useful to input text for this function, for
> example.
> >
> > Unicode is useful but it solves only very basic part of the multilingual
> > issues. For example, how do you distinguish German "Titel", Dutch
"Titel",
> > and Swedish "Titel"?
> > In this sense, an attribute to explicitly identify a language is
> > essential.
> >
> > Multilingual User Interface is also important. However, in our practical
> > environment, I think there is no good straightforward solution.
> Collaboration
> > by native speakers of many languages on the Net seems to be the best way
> > to develop MLUI for the registry from my experience and perspective...
> >
> > Cheers,
> >
> > -- Shigeo
> >
>