Re: When to publish -12
| Newsgroups | gmane.ietf.calendar |
|---|---|
| Message-ID | <OF18E6C698.1B8F9444-ON85256DAA.00683B67-85256DAA.006D9280@notesdev.ibm.com> |
Doug claimed on 09/23/2003 01:31:12 PM:
> Again I am down to zero CAP items.
Not so fast Doug. There are at least 2 other known big issues still to be
resolved, even if you refuse to track them.
1: Busytime in CAP
2: 'Scoping' concerns raised by Preson in April 2003.
I know for one that I did not agree w/Dougs summary of what he thought
Craig proposed (ie: 'Use existing CMD:CREATE to store VFREEBUSY objects."
) and we never fully resolved the issue before the long argument over
RECURRENCE-ID.
Other smaller issues include:
I see that despite my prior postings about the poor ABNF, some changes
have yet to still be made regarding queryc. According to the ABNF, the
_only_ command that can have a VQUERY component in it is the CREATE
command. This is wrong and was already noted in May/June 2003 but still
remains unfixed.
An analysis of the ABNF shows that queryc is ONLY used in the create-comp
which only occurs for the CREATE command. As such, it is not legal by the
ABNF to have a VQUERY inside a DELETE or SEARCH command. Clearly this
was not intended but its not been corrected yet either. The ABNF for some
commands like CREATE are well structured and have relatively clear request
and reply sections but other commands like DELETE do not. The ABNF needs
to be cleaned up and made consistant.
The text of the CREATE command needs to be reordered a bit to be more
logically correct. Currently, right after the "Formal Defintiion" text is
a block titled "Response:". Further down is the text that describes the
actual "request" (my lable). This seems backwards; the "request" bit of
the command should be described before the "response" bit unless you like
confusing readers. This same ordering check should be done to all
commands.
Lastly, I also see some changes that did not get fully done in CAP-12-c or
-d. Some references to stored queries were removed but not all:
QUERYID - A unique id for a stored query. (Section 8.27)
can be removed along with all of Section 8.27 and any other references to
queryid.
Section 9.6 VQUERY Component needs to be updated a bit more to justt:
9.6 VQUERY Component
Component Name: VQUERY
Purpose: A component to specify what is to be fetched from a CS.
Formal Definition: A "VQUERY" component is defined by the following
notation:
queryc = "BEGIN" ":" "VQUERY" CRLF
queryprop
"END" ":" "VCAR" CRLF
queryprop = 1*(
; the following MUST occur at least once
;
query /
; 'expand' is OPTIONAL but MUST NOT occur
; more than once.
expand /
; the following are OPTIONAL, and MAY occur
; more than once
;
other-props
)
Description: A "VQUERY" contains properties that specify which
properties and components the CS is requested to return during a
SEARCH command.
For examples, see Section 6.1.1.
Also, technically speaking a VQUERY is NOT just used "to specify what is
to be _fetched_" because the VQUERY can be (once we fix it) used in more
than the just the SEARCH command. When used on some commands it "specifys
a set of components to apply the command to", whatever the command is that
uses it. That is, it specifys the set of components to delete on the
DELETE command, those to move with the MOVE command, etc. Not every
command "fetches" components so this Purpose is incorrect as it stands
now.
Bruce
===========================================================================
Bruce Kahn INet:
[email protected]
Messaging & Collaboration Phone: 978.399.6496
IBM Software Group FAX: and nothing but the FAX...
Standard disclaimers apply, even where prohibited by law...