[j-nsp] BGP full mesh or route reflector

[Saku Ytti]

On Wed, 21 Jan 2026 at 15:43, James Bensley <lists+junipernsp at bensley.me> wrote:

> I don't think you should start with any assumptions (even if they have been the outcome in 95% of cases).

Writing one line of code can easily be dozen decisions.

We assume all the time. And it's the only way to get something actually done.

If we'd carefully consider every possible question, it would take
prohibitively long to do anything and it still wouldn't impact the
probability of success in a meaningful way.

We hope that the people involved are making a lot of good assumptions.
But now we've decided if that is too slow, to have subject matter
experts make assumptions, so we've spent a few trillion on an
automated assumption machine, using output as it, with no one involved
competent to evaluate it.
-- 
  ++ytti