Re: thoughts on documentation reuse.

[Wes Hardaker <wes@hardakers.net>]

>>>>> On Sat, 23 Mar 2002 16:57:54 -0800, Andy Bierman <abierman@cisco.com> said:

Andy> are you volunteering? :-)

No, I volunteer the libsmi authors ;-)

Actually, I might be willing to throw together a "should work in most
cases" perl regexp based implementation (it shouldn't take much work I
don't think, as you don't have to parse it and look for errors.  Only
compile it into something else which should be then checked via a real
tool.  IE, the compiler wouldn't check for errors, it would only
compile to produce the expanded version.  That should be easy to
implement if you don't have to do error checking ;-) ).

Andy> We need to rethink our TC procedures as they become TYPEDEFs.
Andy> Steve Waldbusser suggested in the RMONMIB WG meeting that maybe
Andy> IANA should maintain common TC modules.

That's a good idea too, though it doesn't help the enterprises case all
the time.  There is always going to be needed TCs that have to be
imported from some odd place.  One huge IANA document would be an
improvement, but people still won't want to look at it half the time
either ;-)

Andy> FooTypeV1 would be the initial version of the type.

hmm..  I need to ponder that notion.  It doesn't sit well initially,
but that may not mean much.

Andy> 2) Relocation
Andy> It's not always easy to recognize that a TYPEDEF is defined in the
Andy> wrong place (e.g., RMON OwnerString).  It might be useful to have
Andy> a RELOCATED clause in the TYPEDEF, to let humans and machines
Andy> know the FooType in the FOO-MIB is the same FooType in a IANA
Andy> common type module.

That's always a hard problem and one that needs to be solved.  It's
really easy to think that "your" module will be the only module that
could ever have a use for such a TC, but invariably at some point
someone is bound to prove you wrong.

-- 
Wes Hardaker
NAI Labs
Network Associates