Re: create-subscription vs. subscribe

[Andy Bierman <ietf@andybierman.com>]

Phil Shafer wrote:
> Andy Bierman writes:
> > Actually, <get-notifications> is probably better because
> > it covers subscribe and replay modes.
>
> But it's not <get>ting them, since they aren't returned immediately.
> It's <start>ing them, or <subscribe>ing to them.  <start> is odd
> too, since there's no <end>, so I'd lean toward <subscribe>, even
> though it's not really a pub/sub model anymore.
>
> And I guess that's the thing that still bugs me, that it's like a
> pub/sub model without the pub/sub.

I agree, but many others think the 'subscribe' model is the
most familiar.  I am concerned that we might be setting up
expectations of feature complexity with this terminology.
(Next year will we have to add the 'publish' half of the model
because that's what is expected when this terminology is used?)

I ran into the same thing wrt/ the term 'transaction'
as used within NETCONF vs. what it means in database terminology.

The verb <subscribe> is fine with me.  Maybe someday a <publish> verb
will be added (but not today ;-).


> I keep thinking that we had two drafts at the interim meeting, one
> simple with few features, and one complex with many features.  But
> it feels like we walked away with a mixture that has most of the
> complexity of the complex one combined with most of the features
> of the simple one.  I know that last afternoon of the interim was
> a whirlwind, but I'm worried we somehow lost the better parts of
> both drafts.

(This is why I wanted you to co-author the RFC. !!!)
Specifically, what needs to be changed?

I noticed that (although undocumented) the <get-syslog-streams> RPC
had a lot more detail in it, such as the <recording/> attribute.
The current stream element doesn't have much in it at all.
How about a list of attributes we should have?

Some changes were made at the WG meeting, and draft minutes
will be done this week.  The 'stopTime' parameter was put back,
and the 'filter' AND 'namedProfile' has been changed to
a choice (OR instead of AND).

> Thanks,
>  Phil

Andy




--
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/>