opaque docs [Re: rough consensus of what "population"?]

[Brian E Carpenter]

I really think MIBs are a special case. As Harald noted, MIBs can be very
instructive even without going through the entire details, and as Pekka
noted, they are frightening. I suspect they will always need special
handling compared with other documents.

Personally, I found that it was the MIB that made diffserv concrete, rather
than any of the other standards track documents. Only in the MIB did we
need to be precise about exactly what parameters have to be set and
monitored in a router to implement and monitor diffserv.

   Brian

Pekka Savola wrote:
> 
> On Tue, 25 Mar 2003, Thomas Narten wrote:
> > > One thing that I was about to raise in that particular w.g. was whether
> > > the majority of the IPv6 wg. is actually *capable* of being able to do
> > > any useful reviewing.
> >
> > Not to pick on you Pekka (you're the messenger here), but a lot of
> > people seem to be awfully proud that they don't know anything about
> > MIBs and want to keep it that way. MIB work is for MIB doctors, seems
> > to be widely held view.
> >
> > But think about the implication of this view. Most people seem to
> > agree that MIBs are important. They get used. They are important to
> > operations. WGs even agree that the work needs to be done. Yet,
> > actually doing and reviewing MIBs seems to be somebody else's job.
> >
> > What is wrong with this picture?
> 
> Yes, I agree it is a problem -- there is no way specialized,
> area-independent MIB doctors could do all the work on MIB's.
> 
> A fundamental problem may be that MIB's look a bit scary.  A formal data
> structure and model is not for the faint of heart.  So, it's something
> that, until you get started (do a few reviews of them at least), is not
> something you look forward to and subsequently want to avoid.
> 
> So, it seems like folks expect that there are some decently-MIB-clueful
> w.g. members, and expect them to do the thing.  Or so it seems..
> 
> I don't know whether/how this can be fixed.  Perhaps if the descriptive
> parts and data structures would be separate that might help -- but then
> again, maybe not.
> 
> Sigh.
> 
> --
> Pekka Savola                 "You each name yourselves king, yet the
> Netcore Oy                    kingdom bleeds."
> Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings