Date:         Wed, 4 Feb 2004 19:40:16 -0500
Reply-To:     "Z39.50 Next-Generation Initiative" <[log in to unmask]>
Sender:       "Z39.50 Next-Generation Initiative" <[log in to unmask]>
From:         "LeVan,Ralph" <[log in to unmask]>
Subject:      Re: XPath and Validation
Comments: To: "Z39. 50 Next-Generation (E-mail)" <[log in to unmask]>
Content-Type: text/plain; charset="iso-8859-1"

> -----Original Message-----
> From: Theo van Veen [mailto:[log in to unmask]]
> Sent: Wednesday, February 04, 2004 7:10 PM
>
> With the current approach people will start putting
> all kind of non-dc garbage in the dc-record because
> that will be the only schema they recognize from
> explain.

Clients are the ones that read explain records and they don't put anything
into dc records.

If you meant the servers, I don't understand that claim.  If they use our DC
schema, then there is no place to put unknown elements; implementors are
restricted to the DC elements we've specified.

Now, non-DC data can be stuck into DC records in the same way that I've been
putting non-Marc-21 data into Marc-21 records.  You add a dc:description
field with content like: "Math Formula: f(x)=x+1".  (I'd have stuck that
same content into a 650 field in a Marc-21 record in classic Z39.50 because
Marc-21 is the only widely interoperable syntax available.)

Content providers who have gone to all the trouble to putting their data
inside a searchable database and then provided an SRU interface to that data
are motivated to make that data available in a usable way.  Dumb clients
will only be able to handle those minimal records.  Users with an incentive
to access the richer records that the provider has will see to it that their
client software can handle a richer schema.

I'm just not seeing a real-world problem here.

Ralph