Re: Issue with MIB compilation requirement in "AD Review of I-Ds" (http://www.ietf.org/ID-nits.html)

[Frank Strauss <strauss@ibr.cs.tu-bs.de>]

Hi!

smilint emits messages at different levels (more fine grained than
just error and warning). A severity up to level 2 is regarded as an
error, i.e. a violation of the standard. A severity with a level in the
range from 3 to 6 (max) is regarded as a warning (but still at several
levels). We can roughly say that the lower half of the warning-level
range is more based on SHOULD NOTs in the specs, while others are based
on experience or consequences that might result from a given construct.

I think, this makes sense. E.g., lint also gives hints on C constructs
that "smell unusual" and let the user have a closer look on what he
wrote. In the end, it is the task of the user to make a dicision. The
same is true for smilint: Errors must be fixed, warnings (at different
levels) are hints to the user.

What is missing, is probably a careful reclassification of the known
warning levels and maybe a more verbose description of what some
messages mean or what the consequences could be.

In case of the "[5] identifier `%s' differs from `%s' only in case"
message, the reason is that a MIB could be mapped to target (programming)
language where identifiers are not case sensitive, so that this would
cause problems. Note that this is a level 5 warning, hence a very low
severity.

 -frank