On 19/10/2017 12:21, Nigel Titley wrote:
I'm generally against additional complication, especially where past practice doesn't give cause to worry, but as I say it doesn't really bother me.
For what the opinion of one of the initial architects of this is worth: Complication and over-specification bothers me greatly. Nothing good has ever come of it. A lot of headaches and some real badness have. Any unnecessary formalism creates friction losses at the very least. I encourage everyone proposing additional formalism to first state very clearly the concrete *need* for adding it and to provide examples of concrete instances where the absence of such formalism has caused problems. Speculative instances in the future only count if there is consensus that they are either very likely to occur or have catastrophic consequences. In the latter case additional scrutiny of whether the added formalism will actually prevent the catastrophe is required. Repeat: state a *need* not a desire or other lesser reason. Daniel speaking as co-founder of RIPE, initial architect of the RIPE NCC association, steady contributor to both and *not* speaking as a RIPE NCC employee