[IPv6]Re: Working Group Last Call draft-ietf-6man-zone-ui

[Erik Auerswald <[e-mail]>]

Hi Brian,

On Thu, Sep 26, 2024 at 08:06:12AM +1200, Brian E Carpenter wrote:
> On 25-Sep-24 23:04, Erik Auerswald wrote:
> >Hi,
> >
> >On Wed, Sep 25, 2024 at 07:08:32AM +1000, Jen Linkova wrote:
> 
> ...
> 
> >I find the length example from section 4 problematic.
> >Systems different from Linux use longer interface names, e.g.,
> >TwentyfiveGigabitEthernet1/0/2 could be an interface name of a Cisco
> >Catalyst switch that can be used as a zone identifier for, e.g., ping.
> >Giving a concrete number, as done in draft-ietf-6man-zone-ui-03, might
> >be more problematic than helpful.
> 
> I don't quite understand your objection. On Linux, a limit of 16 would
> be appropriate. On a Catalyst, some other limit (presumably greater
> than 30) would be appropriate.
> 
> Would it help if the text said:
> 
> "A UI implementation should apply an appropriate length limit..."
> 
> instead of "reasonable", which is certainly vague?

Here, vague is better than giving an exact number.  Otherwise implementers
might just read the "for example, a limit to 16 ASCII characters" part
and conclude that exact limit is always appropriate.  They might consider
"16" to be the standardized number.

Both "reasonable" and "appropriate" work for me.

IMHO just leaving out the example would improve the text in this case.

Br,
Erik