- From: Rich Salz <rsalz@zolera.com>
- Date: Mon, 11 Feb 2002 21:24:09 -0500
- To: Christopher Ferris <chris.ferris@sun.com>
- CC: xml-dist-app@w3.org
I didn't say it IS an issue, I just said that the difference in processing models is likely to cause issues as developers have to straddle both worlds. For example, in [1] the reference you cited, the Infoset says characters are ISO 10646 character codes, "in the range 0 to #x10FFFF", while crypto only uses 8bit bytes. I'm also concerned about things like PSVI and "values applied from DTD declarations" adding data which isn't in the XML serialization used by the node hashing something, e.g. Again, I'm not saying there ARE problems, just that it's possible. /r$ [1] http://www.w3.org/TR/xml-infoset/#infoitem.character -- Zolera Systems, Securing web services (XML, SOAP, Signatures, Encryption) http://www.zolera.com
Received on Monday, 11 February 2002 21:23:59 UTC