W3C home > Mailing lists > Public > www-dom@w3.org > January to March 2004

L3 Core: DOMConfiguration schema-location and schema-type parameters

From: Curt Arnold <carnold@houston.rr.com>
Date: Sat, 10 Jan 2004 23:05:25 -0600
Message-ID: <4000D995.4070108@houston.rr.com>
To: www-dom@w3.org

http://www.w3.org/2004/01/CR-DOM-Level-3-Core-20040108/core#parameter-schema-location

    Represent a DOMString object containing a list of URIs, separated by
    whitespaces (characters matching the nonterminal production S
    defined in section 2.3 [XML 1.0]), that represents the schemas
    against which validation should occur, i.e. the current schema.

Pretty awkward sentence, especially the use of "represent" which also is 
used in schema-type's description.  The i.e. is not helpful.  It is 
unclear if the URI's are namespace/location pairs like 
xsi:schemaLocation or a location list.  Since this is attempting to 
support DTD's, I assume the value is a list of locations.

My interpretation of the next sentence is that xsi:schemaLocation URI's 
in the source document will be retrieved if a qualified name is 
encountered that is not defined in the provided schemas.  If you wanted 
to perform schema validation but wanted to prevent retrieval of schemas 
specified using xsi:schemaLocation, you would need to provide a resource 
resolver that returned an "empty" LSInput.  I would think that ignoring 
schema locations in the instance document if schema-location was 
specified would have been preferrable.

    For namespace aware schema, if a schema specified using this
    property and a schema specified in the document instance (i.e. using
    the schemaLocation attribute) in a schema document (i.e. using
    schema import mechanisms) share the same targetNamespace, the schema
    specified by the user using this property will be used.

If you take out the parentheticals, you end up with this phrase:

    a schema specified in the document instance in a schema document share

which is gibberish.

    It is strongly recommended that Document.documentURI will be set so
    that an implementation can successfully resolve any external
    entities referenced.

This sentence is confusing, "will be" probably should be "be" and 
"external entities" perhaps "relative URIs".  If a schema contains an 
external entity, its location would be relative to the schema's URI.  
Wouldn't it be better to strongly recommend not using relative URI's for 
schema-location?

schema-type parameter

    Represent a DOMString object containing an absolute URI and
    representing the type of the schema language used to validate a
    document against.

Perhaps:

    A DOMString containing an absolute URI specifying the schema
    language used in validation.
Received on Sunday, 11 January 2004 00:05:29 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 22 June 2012 06:13:57 GMT