<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body>
<div class="moz-cite-prefix">On 2023-09-08 09:54, Rhys Barrie via
ARIN-consult wrote:<br>
</div>
<blockquote type="cite"
cite="mid:CALQYBjJiu2i6d11_zVeamhL3_v840wEEVD19nLOJ7AO4apGTRA@mail.gmail.com">
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
<div dir="ltr">In my opinion, opt-in is functionally equivalent to
not changing anything at <br>
<div>all, because 95% of organizations will never make the
conscious effort to click <br>
the button.</div>
</div>
</blockquote>
<p>Agreed.<br>
</p>
<br>
<blockquote type="cite"
cite="mid:CALQYBjJiu2i6d11_zVeamhL3_v840wEEVD19nLOJ7AO4apGTRA@mail.gmail.com">
<div dir="ltr">
<div>hearing Job's feedback, and seeing LACNIC's success with a
<br>
tightly-coupled integration, I am reasonably convinced as to
the validity and <br>
efficacy of that solution, and I agree that opening the door
to discrepancies <br>
WILLÂ cause discrepancies (and outright issues as a result) to
occur. I believe <br>
that pure opt-in/opt-out exacerbates the long-tail problem as
well.<br>
</div>
</div>
</blockquote>
<p>Agreed.</p>
<p><br>
</p>
<p>As long as we are talking about RPKI automating IRR, this should
be fine.</p>
<p>The opposite (creating a route object creates a ROA) is not
reasonable, IMHO. I don't think anyone has suggested that
approach, though.</p>
<p>Also, AFAIK, nobody has suggested that IRR route objects would
require ROAs, right? As of right now, I still have the problem
that I can create IRR records, but not ROAs, as ROAs can only be
created for direct allocations, not reallocations.</p>
<pre class="moz-signature" cols="72">--
Richard</pre>
</body>
</html>