- From: Scott McGlashan <scott.mcglashan@pipebeach.com>
- Date: Tue, 23 Jul 2002 12:25:49 +0200
- To: "Mukul Jain" <mujain@cisco.com>, "Www-Voice@W3. Org" <www-voice@w3.org>
Hi Mukul, thanks for your comments. You are correct in assumption about the definition of <object>: since objects are platform-specific, the platform itself needs to define exactly what values should be assigned to the object parameters. 1. codetype/type could be a Java applicaiton for example. 2. codebase indicates the general base URI used to resolve relative URIs in classid, data and archive. Classid is a URI for the object's implementation -- if relative, it is resolved against the codebase. 3. the assumption is that there is one master class associated with the object --- it in turn can load other classes following standard procedures in java, c++, etc. I hope this helps. If you think there is a specific problem with the specification which prevents its proper implementation and use, then please let me know. If so, please suggest precise wording for any proposed alternatives. thanks Scott [leader, W3C VBWG VoiceXML 2.0 dialog team] -----Original Message----- From: Mukul Jain [mailto:mujain@cisco.com] Sent: Tuesday, July 02, 2002 21:31 To: Www-Voice@W3. Org Subject: Clarification on <OBJECT> element details. Hi, I feel the <object> element definition is pretty loose and subject to readers interpretation. It could be intentional as <object> elements are always going to be platform specific anyway. Seeking clarifications for: 1.. What could be the possible values for "codetype" and "type" attributes say for example object is implemented using C++ or JAVA? 2. Definition of "codebase" and "classid" is also not very clear as one of them seem to be standard URI convention whereas other one is platform-dependent and then it is said that "codebase" is used to resolve relative path. It would be little counter intuitive to combine a platform-dependent URI convention with one of the standard URI convention for the purpose of resolving relative paths. 3. Does the current form of spec takes care of situation where object implementation spreads in multiple classes? If, yes, how do we define classids for other classes. Thanks, Mukul Jain mujain@cisco.com
Received on Tuesday, 23 July 2002 06:18:32 UTC