From: Pce <[email protected]> on behalf of Dhruv Dhody <[email protected]>
Sent: 23 November 2020 10:44

<tp>
I see no mention of pcep-yang which is one of the hundreds of I-D updated in 
the prelude to the window closing for the meeting:-(  I have looked at it 
briefly.

Pagination; the IETF has abolished it.  Those doing so expect us to have short 
sections that are easy to reference; they are of course unaware of YANG where a 
section can run unbroken for 50 pages.  So, it behoves us to keep sections 
short where that is not impossible.  My take is that a section should be half a 
page or less.  It would be worth looking at the first six sections to see if 
they can sensibly be divided.  I would suggest that wherever a page number is 
missing from the TOC then a change would be good (except for YANG).

There is a lot of 'if pce or pce and pcc'.  I wonder if a YANG doctor could 
find a more elegant expression.

Likewise but more so for me, there is a lot of must have a  value for these 
states or else must be zero with some fairly hard to follow IMHO combinations 
of 'and' 'or' 'not'.  I would have thought it simpler to have a 'when' such 
that the leaf is absent from the model (as opposed to present and MBZ) if the 
states are not suitable.

Again, for me, I find the repeated use of 'domain' confusing.  There is a YANG 
convention of list plural leaf singular so /domains/domain/ as opposed to 
/domain/domain/ but even so there is a lot of domain which I would prefer to 
have slightly differing names.

Finally, is PCE a routing protocol?  There are conventions for routing 
protocols such as OSPF about defining an identity and where to augment which 
this does not follow suggesting that it does not qualify.

Tom Petch


Hi WG,

Please find the minutes for the PCE WG session -
https://datatracker.ietf.org/doc/minutes-109-pce/

_______________________________________________
Pce mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/pce

Reply via email to