If you want to reduce this to analogies, your proposal is much closer to
the situation we had in the IETF a couple of decades ago when it came to
NATs: we preferred architecturally pristine approaches that shattered
into useless, broken shards on contact with reality.
/a
On 9/18/17 17:10, Carsten Bormann wrote:
On Sep 18, 2017, at 23:12, Adam Roach <adam@xxxxxxxxxxx> wrote:
I think "putting it in is demonstrably necessary for proper rendering in a lot of real-world applications" is a Really Good Reason.
That would have been a Really Good Reason to leave security off on the IETF infrastructure.
Security creates so many problems; it’s best to leave it off.
And security for IETF services does not really matter much in practice.
Still, we are doing Security to lead the pack.
It’s all about sending the right signals.
Oh, and we have had countless IPv6 issues, a Really Good Reason to leave that off, too.
What is so disappointing about this discussion is that it’s occurring in 2017.
In 1997, there might have been a point.
On September 3, 1967, road driving in Sweden was switched to driving on the right side.
There was no transition period; people had to run their systems (cars) using the new rules right away.
(It is maybe educating that the popular vote about this change in 1955 was 15.5 % for, 82.9 % against making the change.)
Soupification approaches like BOM pollution are the road to eternal transition.
It feels like we are going to be driving on both sides of the street for the foreseeable future, forcing all right-side drivers to watch out all the time for the remaining left-side drivers, putting in lots of infrastructure to deal with the inevitable collisions, and just generally driving very slowly.
The transition to UTF-8 is pretty much done out there in the real world.
Still, as with Security and IPv6, IETF MUST be leading, not lagging in this space.
It’s all about sending the right signals.
I intend this to be my last message on this thread (but I’m making no promises).
Grüße, Carsten