>>>> [log in to unmask] 09-12-02 18:12 >>>
>> > What is stopping me dumping all of my 1.2 megabyte ead record into
>> > the <dc:description> element?
>>
>> Not that I like Theo's proposal, but -- Rob, this question is one
>> you're going to have to answer whether Theo gets his way or not. What
>
>The fact that you've said it's validatable by the schema at
>http://www.loc.gov/zing/srw/dcschema/v1.0/ and that doesn't mention EAD
>as being valid below <dc:description>
>
Allthough this is not relevant for the original discussion, you do not know what is under dc:description.
>
>> > Au revoir, interoperability.
>> Theo's perspective is that two bits of software, server and client,
>> can be plugged together and produce some kind of result, then you have
>> interoperability -- and so you do, of a kind. (Let's call it
>> "syntactic interoperability"). But you're concerning yourself further
>> with whether you can actually interpret the records you get back.
>
>Of course. It's meaningless to say 'I have interoperability' if I can't
>actually do anything with the record because I don't know what the <title>
>tag in it means or what all these <ead> elements are.
>
You know what title means, but you neglect only a single term that you do not know. So you can still use most of the record. I do not propose to polute the records with all kinds of rubbish.
>>
>> ``This is a Dublin Core record, but I've polluted it with some extra
>> elements from the "rec" namespace. Use them if you understand them,
>> ignore them otherwise."
>
>Which we can do by defining a trivial schema which imports DC and REC.
>
Yes, but how do I know that you did that? I will allways ask for DC because I do not know you have a richer record, and I am not going to search the whole world for new schemas.
Theo
|