I've thought about/messed with this stuff before, and come up with no 
good elegant solution. It is indeed kind of a mess.

On 2/24/2011 12:03 PM, Robert Sanderson wrote:
> That is (still) incorrect.
>
> A single schema may contain multiple namespaces, and there isn't a
> unique identifier for a schema.  For example, any simple Dublin Core
> based syntax must have at least two Namespaces, Dublin Core and the
> wrapper element. SchemaLocation is not unique as there can be many
> copies of the same schema.  A single schema may define multiple root
> elements, such as MODS does with both item and collection level
> elements.
>
> Referring to your blog post, you can say how the four inter-relate:
>
> Schema Identifier uniquely identifies the format.
> Schema Location is a non-unique description of the format.
> Schema Name is a short, human readable, non-unique name for the format
> and Namespace is a non-unique namespace used by the format.
>
> This is just a rehash of a previous discussion on this list, between us:
>
> [log in to unmask]" target="_blank">http:[log in to unmask]
>
> So I guess I'm wasting my time ;)
>
> Rob Sanderson
>
> On Thu, Feb 24, 2011 at 9:44 AM, Jakob Voss<[log in to unmask]>  wrote:
>> Hi,
>>
>> We are developing a general API management tool to provide different APIs
>> (unAPI, SRU, OAI-PMH...) with different record formats (MARC, MODS, DC...)
>> to our databases. We now stumbled upon some confusion regarding XML formats.
>> The basic question is "what is a format and how do you refer to it?"
>>
>> I came to the conclusion that at least SRU schema identifiers are useless.
>> In addition you can extract XML namespace URIs from XML Schemas, so all you
>> need to identify a format is a link to its XML Schema.
>>
>> I wrote a more detailed blog posting about this at
>> http://jakoblog.de/2011/02/24/xml-schema-vs-library-apis-oai-pmhsruunapi/
>>
>> Does anyone of you relies on SRU schema identifiers when consuming SRU?
>> I think at least for XML-based formats we should only use the XML Schema as
>> authoritative reference. Sure there are different applications of variants
>> of one schema, but then it makes no sense to use global identifiers in
>> addition to local names.
>>
>> Jakob
>>
>> --
>> Jakob Voß<[log in to unmask]>, skype: nichtich
>> Verbundzentrale des GBV (VZG) / Common Library Network
>> Platz der Goettinger Sieben 1, 37073 Göttingen, Germany
>> +49 (0)551 39-10242, http://www.gbv.de
>>