Comments on Netconf Issue List

["Sharon Chisholm" <schishol@nortelnetworks.com>]

Hi

Here are some comments I've received on the Netconf issues list. Hopefully
by putting them all into one email I can be forgiven for not checking to see
if they are been resolved in the 02 version or if the specific comments have
already been made.

1) 5.1) End of message directive

On a related note, need clarification on 
what is the convened upon reaction if a syntax error is detected in the XML
input; abort and drop everything?  resynch on the next eom marker?
something configurable through capability exchange?.

2) 10.3.2) (NEW) Confirmed commit

No text below describing this one.  If it means a new command to clearly
indicate the confirm on a commit-confirm sequence, then we like it

3) 12.6) <rpc-error> 

We need a  definition that indicates if the blocks 
can be repeated or not and how they can occur in a hierarchical 
output logic. Also looking for a clear end of response indication with
success status and time-stamp 

4) 13.1) <edit-session>

Can't say from the text if that means edit-session is there or abandoned...

5) 13.3.3) No format parameter for text or XML encoding

It might be worth considering the pros and cons of using namespace here
versus clear parameterization.

6) 13.4) <get-config>

There are implementation issues related to the fact there has been no
discussions on the need for limited meta-model identifications (i.e.
distinguishing between components, attributes, and artificial container XML
elements).  If the only way to use 
this macro language is to throw a huge XML infrastructure in front of it, 
then it will be harder to convince all these people writing plain CLI macros
to use it.

7) 13.8) <get-all>

It would be good if this resulted in a redefinition of the message.  As
defined, 
get-state is basically a generic passthrough-command wrapper and 
an operational data extraction.  With a name like get-all, seems to me that
we are imposing a semantic meaning to this.  I don't see why this command
would necessarily get ALL stuff.  Get all of what anyways... 

Sharon Chisholm
Portfolio Integration
Nortel Networks
Ottawa, Canada

--
to unsubscribe send a message to netconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/netconf/>