Re: [babel] [Last-Call] RtgDir review: draft-ietf-babel-information-model-11.txt

[Mahesh Jethanandani <[e-mail]>]

Hi Geoff,

Appreciate your review and pointing out the nits as you see them.

> On Oct 20, 2020, at 9:36 AM, Geoff Huston <[e-mail]> wrote:
> 
> Why define some intergers as unsigned 16 bit values yet leave strings of undefined maximum length. Can I stuff a novel into the field? How about all of wikipedia? The point is why is the document inconsistent about its treatment opf the range of acceptable values in terms of size?

The reason some variables have fixed length by virtue of the fact that they are unsigned int or unsigned short is because we have a better sense of what the range of those values are. For example, you do not want the babel-mcast-hello-interval to be more than 32 bit long. The same is not true for something like babel-interface-metric-algorithm, which is a name. While we could artificially limit the string value to say 40 characters, what is to say that someone cannot come up with a name that is 41 characters long.

Implementations of this information model, whether they are YANG data model or some other modeling language can choose to further restrict the size of the string. But even there, I have seen YANG models deferring it to particular implementations to refine the definition by adding a size attribute to the node. At an information model level we are poor judges of what size string is ok for any given variable.

Thanks.

Mahesh Jethanandani
[e-mail]