Re: [netmod] draft-ietf-netmod-rfc6087bis-13 - section 4.26.2

[Martin Bjorklund <[e-mail]>]

Hi,

"Bogaert, Bart (Nokia - BE/Antwerp)" <[e-mail]> wrote:
> I would like to understand why the YANG 1.1 feature logic is *much more
> expensive* than YANG 1.0.

The document says "much more _expressive_".

> As far as I can see the way YANG features are
> being defined has not changed between YANG 1.0 and YANG 1.1.

if-feature can in YANG 1.1 be a boolean expression of features, e.g.,
"foo or bar and not baz".


> On the other hand, the second paragraph of this section seems to deal with
> "when" versus "if-feature" and the preference to use if-feature instead of
> when, if possible.  But as far as I'm aware there are no changes w.r.t. when
> between YANG 1.0 and YANG 1.1.  This paragraph seems to suggest that "when"
> is worse than if-feature.  I can understand that when is to be evaluated and
> depends on the when-condition while a feature can be considered as a design
> and implementation choice (the feature is supported or not) and does not
> need any run-time 'validation'.  But why is this so different in YANG 1.1
> versus YANG 1.0?

This was true also in 1.0 (i.e., 'if-feature' is simpler than 'when')



/martin




> Where can we find more background on the statement made in this section
> about much more expensive and what exactly is meant by this, certainly when
> we want to see this in the perspective of the run-time characteristics and
> impact on a NC server running in a device.
> 
>  
> 
> Thanks in advance,
> 
> Bart
>