I apologize for the delay in following up on this thread.
Going back to the four original questions, here are the WiFi team's
decisions and rationale:
1. Should -p be needed to specify properties to the link property
subcommands?
Yes. Even though it is a little more typing, it:
* Is recommended by CLIP.
* Is used by recent utilities, such as svcprop.
* Is more "future-proof" to changes to dladm.
* Sidesteps the "getent operand ordering" problem.
2. What should the link property subcommands be called?
{set,reset,show}-linkprop. The rationale is:
* These names make it explicit about what type of property is
being operated on.
* There is existing precedent (as per David-John's comments).
3. Should the link be specified as an operand to the link property
subcommands?
Yes. As per David-John's comments, usability studies have shown that
having an object sometimes be an operand and sometimes an option
(within the same command) causes confusion. As usual with usability,
it seems that consistency is paramount.
4) What should the secure property subcommands be called?
Many things were proposed, but no consenus was reached here for
something markedly better. One possibility that was brought up
offline was to call these "secure objects", and use *-secobj; this
seems the closest to the intended semantics. If folks like that,
we will use it. Otherwise, we will go with the current *-secprop
names.
--
meem
_______________________________________________
networking-discuss mailing list
[email protected]
