RE: Resolution of open issues: JW24c (DAV:eq)

> From: www-webdav-dasl-request@w3.org
> [mailto:www-webdav-dasl-request@w3.org]On Behalf Of Amelia Carlson
> Sent: Wednesday, February 20, 2002 12:30 AM
> To: Jim Davis
> Cc: Julian Reschke; dasl
> Subject: Re: Resolution of open issues: JW24c (DAV:eq)
>
>
>
> Hello, and greetings, I am new to this 'group'.
>
> For character comparison across languages, probably they
> safest/most well defined solution would be to say they are mapped
> to Unicode and compared there.

From a protocol point of view, properties always are marshalled using XML,
so by definition the server receives them in Unicode (after XML parsing).
Some exceptions are properties derived from HTTP headers (content-type), but
those should be in ASCII anyway.

> However, note that if you specify this only happens when the languages
> differ,
> you can get different results in less than and greater than comparisons if
> the
> Unicode ordering of characters differs from the underlying language when
> all you have is a dialect difference.  Since this issue is specific to
> DAV:eq,
> perhaps that is not a concern.
>
> I've seen cases where the ordering of characters differs from the
> language-specific ordering to Unicode, but am not familiar with those
> languages enough to know if they had multiple dialects available.

My concern is that the WebDAV server may not have control about what
encoding/collation sequences the backend database uses. So *requiring* some
specific behaviour is likely to make the spec not implementable.

DAV:basicsearch should stay simple -- otherwise we'll never get it out of
the door...

Julian

Received on Wednesday, 20 February 2002 08:40:08 UTC