[art] Re: [Uri-review] Re: Alternative representation of URIs in YANG

[Mark Nottingham <[e-mail]>]

I understand that from a YANG perspective the focus is on getting something configured. However, what's being talked about here is creating a protocol artefact that claims to be "ietf-uri" -- i.e., something that the whole IETF community considers to be a good way to convey a URI -- and yet there is clear and consistent feedback from the relevant expert community that what is described is not an appropriate construct. 

As has been pointed out by others, handling and comparing URIs requires a _lot_ of subtlety, compounded by the many different syntactic options that they offer. It's not at all clear that this structure has been thought through with reference to how it will be used -- hence the nervousness you're hearing from the URI expert community.

Furthermore, if something is called "ietf-uri", it is likely to be reused in other situations where the constraints you have in mind are no longer applicable. 

So the fairly strong advice you're getting is to either:

a. Just use a string, per MT (preferred)
b. Rename to something *much* more specific to your use case

Cheers,


> On 24 Jan 2026, at 9:56 am, Mahesh Jethanandani <[e-mail]> wrote:
> 
> Hi Henry,
> 
> Can you provide a specific example of how the current definition of 'ietf-uri’ in YANG does not meet the requirements of configuring HTTP?
> 
> Thanks.
> 
>> On Jan 23, 2026, at 4:22 AM, Henry S. Thompson <[e-mail]> wrote:
>> 
>> Mahesh Jethanandani writes:
>> 
>>> How about we go with option 2 that limits the applicability of the definition to the following:
>>> 
>>> The 'ietf-url' module defines a YANG 'grouping' for a URL described as a
>>> constrained subset of the URI defined in <relref section="3" target="RFC3986"/>.
>> 
>> Just changing the 'i' to a 'u' doesn't change the fact that it's no
>> longer a UR[IL] if it's restricted to 'HTTP', so IMHO it's still
>> misleading.
>> 
>> ht
> 
> 
> Mahesh Jethanandani
> [e-mail]
> 
> 
> 
> 
> 
> 
> _______________________________________________
> art mailing list -- [e-mail]
> To unsubscribe send an email to [e-mail]

--
Mark Nottingham   https://www.mnot.net/