In the earlier version and in this version I note that there are a number of header fields that do not declare in which messages they may appear. The typical statement (e.g. as for Weight) is something like This header MAY occur in START-PHRASE-ENROLLMENT and MODIFY-PHRASE requests There is no such specification for a number of the speech service-specific messages including: Save-Best-Waverform Recognition-Mode Personal-Grammar-URI Also with the Generic Message Header fields there is no similar strategy employed. With the exception of certain "conditional" headers, like Active-Request-Id-List for STOP messages, it would appear that most are optional for all message types, yet this is not stated. Would there be any benefit in being more explicit about when header fields should occur or not - maybe in some tabular form? For example, I have attempted to perform an analysis or the presence of header fields across some message types... See the attached spreadsheet. You will see where I am having difficulty in determining whether a header field should be present or not and when it is what might be the potential impact. Kind Regards, Arthur Vernon Lead Software Architect IST Networks Unit 5B, 1 Station St, Subiaco WA 6008 AUSTRALIA (m) +61 411 336 176 (w) +61 (8) 6380 2058
Attachment:
MRCP.xls
Description: MRCP.xls
_______________________________________________ Speechsc mailing list Speechsc at ietf.org https://www.ietf.org/mailman/listinfo/speechsc Supplemental web site: <http://www.standardstrack.com/ietf/speechsc>