- From: Noah Mendelsohn <noah_mendelsohn@us.ibm.com>
- Date: Mon, 10 Dec 2001 13:06:24 -0500
- To: "Jacek Kopecky <jacek" <jacek@systinet.com>
- Cc: "xml-dist-app" <xml-dist-app@w3.org>
As I've suggested before, I think there are issues relating to external references that go beyond the encodings, and I think our approach has to be consistent across the cases. If I send you a document that uses encoding and has an href to some other URL, what are my obligations in following that link? It has very bad performance and security implications if you even imply that a conforming implementation MUST try to open a random URL that happens to show up in the href of a document. This also relates to our handling of SOAP+Attachments and DIME, which I think we've delayed for now. So, we need to indicate in the encodings, what is the result if there is an href you choose not to follow or can't follow? Is it that a fault should be generated? Is it the same fault as if an href in the form of a fragment referencing the envelope itself fails? Thanks. ------------------------------------------------------------------------ Noah Mendelsohn Voice: 1-617-693-4036 Lotus Development Corp. Fax: 1-617-693-8676 One Rogers Street Cambridge, MA 02142 ------------------------------------------------------------------------ Jacek Kopecky <jacek@systinet. To: <xml-dist-app@w3.org> com> cc: Sent by: Subject: issue 168 proposal: xsi:type of external references in Encoding xml-dist-app-req uest@w3.org 12/09/01 02:05 PM Hi all. 8-) This is a proposal for resolving issue 168 [1]. The issue wants to know the XML Schema Instance type of the data that is referenced by the href attribute and that is external to the referencing document. The proposal: The type of referenced data depends on whether the referenced document is an XML document or not. The possible means of determining whether a particular document is XML or not include using MIME types, file extensions or file introspection. In case the document is XML, the referenced data is either the root element or the element identified by its ID (in case the referencing URL contains the #<id> part), the type of this data is determined the same way as if the reference was internal. In case the document is not XML, the type of the referenced data is enc:binary, a simple type with the value space of the set of finite-length sequences of binary octets and the same lexical space. [guess I'll need somebody's help with this definition] A different proposal could try to introduce some means of otherwise typing the data (for example an attribute enc:hrefType). To me this seems too complicated. Yet another proposal could disallow external references altogether. This would IMO meet big push back from the proponents of the current practice of attaching data. Anyway, here it is, read for being discussed. 8-) Jacek Kopecky Senior Architect, Systinet (formerly Idoox) http://www.systinet.com/ [1] http://www.w3.org/2000/xp/Group/xmlp-issues.html#x168
Received on Monday, 10 December 2001 14:51:44 UTC