Er, conventional wisdom has that the clue density is constant.
The 'net grows. Appearence of clue is reduced.
The CAIDA series is really a good idea. Mentoring is a great idea.
Potentially partnering w/ USENIX once a year would be of great
benefit. ARIN et.al. are trying to get some WG's off the ground.
There, a few good ideas. Tech. transfer, Empirical (SOP) knowledge,
Cross Disipline meetings. All waiting for someone to take charge
and complete the task. Not sure that it can be in the current
NANOG form thou.
I'd like to propose a panel at the October NANOG in which people talk about
what skills they'd like new engineers to come with, what sorts of internal
formal training they do, any mentoring/self-study guidelines, and the
needs, if any, for formal external training. Ideally, this might include
both providers and selected users, the latter dealing with relevant issues
such as working to establish multihoming through less-than-clueful ISPs.
The goal would be a first cut at a roadmap for imparting clue. Perhaps
this might also work as a BOF similar to the one on facilities that Bill
Norton led.
Disclaimer: my firm does, among other things, do training services,
including developing hopefully realistic exterior routing labs. Our major
motivation, however, is that our healthcare networking customers often need
high-availability connectivity, and we've been finding we need to educate
some local ISPs on how to service our customers.
