Re: linux-ipsec: organizational lessons from Raleigh

>>A secondary issue there is that a bit of advance interop testing against a

I'm not surprised, but if we're going to make an effort to provide for dissimilar-implementation testing, we should aim for as much dissimilarity as possible. It's quite conceivable that the common heritage of those two implementations could generate common problems; we should try for testing against at least one implementation with an entirely independent origin (if that can be arranged -- OpenBSD does have the advantage of being free).

Henry Received on Fri Mar 6 23:12:14 1998