<!DOCTYPE html>
<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    <p>Hi,</p>
    <p><br>
    </p>
    <p>Sorry for top-posting, but it seems exactly me cares about most
      of that anyway.</p>
    <p><br>
    </p>
    <p>It's a risk.  We're aware of it.  If you want an object restored
      and your not happy with the outcome, escalate the matter.</p>
    <p><br>
    </p>
    <p>If you're unhappy with the wording, propose alternative wording
      or live with it.  It's that simple.</p>
    <p><br>
    </p>
    <p>The inter-RIR risk is currently also there, so irrelevant, plus,
      I think Afrinic is the only RIR that still allows new "flat"
      AS-SET objects to be created.  And someone pointed out to be today
      RADB, but I haven't confirmed that.</p>
    <p><br>
    </p>
    <p>That also reduces your risk of creating exactly the collision
      that we intend to prevent because those collisions would exist
      currently (as in before the effective date where no more flat
      AS-SET objects can be created)!</p>
    <p><br>
    </p>
    <p>If you can propose better wording, and with enough foresight to
      deal with all possible future restore situations in a sensible way
      I'm sure James would happily reconsider again.  But I doubt that's
      possible bringing us back to even with an extremely lenient, and
      specifically vague motivation for restore we're still
      significantly better of than what we are today.</p>
    <p><br>
    </p>
    <p>Kind regards,<br>
      Jaco</p>
    <p><br>
    </p>
    <div class="moz-cite-prefix">On 2026/07/29 17:11, Tshepo Masuku
      wrote:<br>
    </div>
    <blockquote type="cite"
cite="mid:DB7PR04MB5993CCA0CDBE8B92F146F0A7CACA2@DB7PR04MB5993.eurprd04.prod.outlook.com">
      <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
      <div dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        <br>
      </div>
      <div dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        Dear Jaco,</div>
      <div dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        <br>
      </div>
      <div dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        Thank you for explaining your view. However, your concession
        identifies the problem: if one staff member may approve a
        restoration and another may reject the same request, the rule is
        not deterministic. Whether we call that vague or ambiguous, an
        operator cannot know the outcome from the policy text.</div>
      <div dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        <br>
      </div>
      <div dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        There is also a specific inter-RIR risk. A flat AS-SET may be
        deleted from AFRINIC and the same name may then be created in
        another authoritative IRR before restoration. Restoring the
        AFRINIC object “exactly as it was” could recreate the very
        cross-registry collision this proposal is intended to prevent.
        Confirming that the key remains free inside AFRINIC would not be
        sufficient.</div>
      <div dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        <br>
      </div>
      <div dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        If restoration is retained, the policy should require verified
        prior control, restoration from the archived object rather than
        creation of a different object under the old name, and a
        conflict check at the time of restoration. It should also define
        how a refusal is reviewed.</div>
      <div dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        <br>
      </div>
      <div dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        I do not doubt that AFRINIC staff may act reasonably. But policy
        must survive staff changes and cannot depend on assumptions
        about how generously discretion will be exercised. Trust is not
        a substitute for an objective rule.</div>
      <div dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        <br>
      </div>
      <div dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        My concern is therefore not that the proposal fails to solve
        every IRR problem. It is that its exception can potentially
        recreate the precise problem the mandatory rule claims to close.</div>
      <div dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        <br>
      </div>
      <div dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        Regards,</div>
      <div dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        Tshepo</div>
      <div dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        <br>
      </div>
      <div dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        <br>
      </div>
      <div id="ms-outlook-mobile-body-separator-line"
        data-applydefaultfontstyles="true" dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt;">
        <div
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt;">
          <br>
        </div>
      </div>
      <div id="ms-outlook-mobile-signature" dir="auto"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">
        <span
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(33, 33, 33);">Get
        </span><a href="https://aka.ms/AAb9ysg"
style="font-family: Aptos, Aptos_MSFontService, -apple-system, Roboto, Arial, Helvetica, sans-serif; font-size: 12pt;"
          moz-do-not-send="true">Outlook for Android</a></div>
      <hr style="display:inline-block;width:98%" tabindex="-1">
      <div id="divRplyFwdMsg" dir="ltr"><font face="Calibri, sans-serif"
          style="font-size:11pt" color="#000000"><b>From:</b> Jaco Kroon
          <a class="moz-txt-link-rfc2396E" href="mailto:jaco@uls.co.za"><jaco@uls.co.za></a><br>
          <b>Sent:</b> Wednesday, 29 July 2026 16:48:37<br>
          <b>To:</b> Tshepo Masuku <a class="moz-txt-link-rfc2396E" href="mailto:TshepoMasuku26@hotmail.com"><TshepoMasuku26@hotmail.com></a>;
          James Bensley <a class="moz-txt-link-rfc2396E" href="mailto:james@inter.link"><james@inter.link></a>; <a class="moz-txt-link-abbreviated" href="mailto:rpd@afrinic.net">rpd@afrinic.net</a>
          <a class="moz-txt-link-rfc2396E" href="mailto:rpd@afrinic.net"><rpd@afrinic.net></a>; <a class="moz-txt-link-abbreviated" href="mailto:pdwg-chairs@afrinic.net">pdwg-chairs@afrinic.net</a>
          <a class="moz-txt-link-rfc2396E" href="mailto:pdwg-chairs@afrinic.net"><pdwg-chairs@afrinic.net></a><br>
          <b>Subject:</b> Re: [rpd] [Last Call] Draft Policy Proposal -
          Hierarchical Names for New AS-SETs
          (AFPUB-2026-ASN-001-DRAFT02)</font>
        <div> </div>
      </div>
      <div>
        <p>Hi Tshepo,<br>
          <br>
        </p>
        <div class="x_moz-cite-prefix">On 2026/07/29 15:48, Tshepo
          Masuku wrote:<br>
        </div>
        <blockquote type="cite">
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            Dear James,</div>
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            <br>
          </div>
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            Thank you for considering the concern carefully and
            explaining your decision openly. I genuinely appreciate that
            you engaged with the substance rather than simply dismissing
            it.</div>
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            <br>
          </div>
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            I understand the case for incremental improvement and agree
            that no policy can anticipate every future scenario. My
            concern, however, is not that the draft must be perfect. It
            is that the exception clause sits at the boundary of what
            AFRINIC will be required, or permitted, to enforce.</div>
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            <br>
          </div>
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            The absence of data cuts both ways. If we do not know how
            frequently restoration will be needed, that uncertainty does
            not necessarily justify leaving the criteria undefined. It
            may instead support beginning with a narrow, objective
            restoration rule and expanding it later if operational
            evidence shows that broader exceptions are necessary.</div>
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            <br>
          </div>
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            I also do not think restarting Last Call is, by itself, a
            sufficient reason to retain ambiguous wording. Once adopted,
            the policy will be binding until another proposal completes
            the PDP. During that period, AFRINIC will have to interpret
            “where necessary” without agreed criteria. That is not quite
            the same as a temporary implementation experiment.</div>
        </blockquote>
        I don't think it's ambiguous, it may be vague, but not
        ambiguous.  I will concede it leaves the jay or nay decision to
        operational discretion, and one person may decide yes, another
        may decide no, but IMHO *any* reason to restore an object would
        be operational of nature, something like "Hi, we've deleted
        object X, but it has come to light it's still being referenced
        by {insert another object|operational peer name|AS} and we're
        having trouble getting the relevant party to update, could you
        please restore".<br>
        <br>
        I don't foresee any other case of restoration case that I would
        consider to be valid.  But precisely that "cannot foresee" is
        why I believe it's best to leave it to Afrinic staff to
        determine the validity of any motivation for *restoration* of an
        object.  At that point it becomes a debate.  And I've not yet
        encountered Afrinic staff to be unreasonable and they'll most
        likely err on the side of restoring (exactly as it was, without
        modification) rather than not.
        <blockquote type="cite">
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            <br>
          </div>
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            A narrowly drafted restoration safeguard would not be an
            attempt to build the whole castle at once. It would simply
            make the first brick clear enough that operators and staff
            know what it permits. The common layer should be minimal,
            but what it contains should also be deterministic and
            auditable.</div>
        </blockquote>
        <p>Without knowing what will or may be required in the future
          this is exceptionally difficult, as per James, to narrow
          down.  I agree with James that it's best to intentionally
          leave it vague.</p>
        <p><br>
        </p>
        <p>Bottom line is:  no new "flat" objects, and restore only on
          motivation to Afrinic staff (which will likely be in the lines
          above).<br>
          <br>
          At some point we would likely want to get a list of all "flat"
          objects and see if they're still *referenced*, confirming
          actual *use* is harder since we have no way of knowing whether
          any implementation references that object.  At least, none I'm
          aware of.  If there is a way it would likely involve Afrinic
          having to check some query log and if the object name hasn't
          been queried in some time.  But that's not to say there aren't
          gateway systems (such as rr.ntt.net - default queried server
          by at least bgpq4) running some form of irrd, that doesn't
          pull the entire database and then references "offline", so
          without looking into exactly how that infrastructure works,
          there's no way to identify completely unused objects I don't
          think.</p>
        <p><br>
        </p>
        <p>I trust this answers your concern.</p>
        <p><br>
        </p>
        <p>Kind regards,<br>
          Jaco</p>
        <p><br>
        </p>
        <blockquote type="cite">
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            <br>
          </div>
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            I respect your decision not to revise the draft, and I
            appreciate your constructive approach. However, because the
            difference between a guaranteed restoration right and
            discretionary case-by-case approval remains unresolved, I
            cannot support the proposal as currently written and
            therefore maintain my objection.</div>
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            <br>
          </div>
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            kind regards,</div>
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            Tshepo</div>
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            <br>
          </div>
          <div dir="auto"
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
            <br>
          </div>
          <hr tabindex="-1" style="display:inline-block; width:98%">
          <div id="x_divRplyFwdMsg" dir="ltr"><font
              face="Calibri, sans-serif" color="#000000"
              style="font-size:11pt"><b>From:</b> James Bensley
              <a class="x_moz-txt-link-rfc2396E"
                href="mailto:james@inter.link" moz-do-not-send="true"><james@inter.link></a><br>
              <b>Sent:</b> Wednesday, 29 July 2026 14:26:30<br>
              <b>To:</b> Tshepo Masuku <a
                class="x_moz-txt-link-rfc2396E"
                href="mailto:TshepoMasuku26@hotmail.com"
                moz-do-not-send="true">
                <TshepoMasuku26@hotmail.com></a>; <a
                class="x_moz-txt-link-abbreviated moz-txt-link-freetext"
                href="mailto:rpd@afrinic.net" moz-do-not-send="true">
                rpd@afrinic.net</a> <a class="x_moz-txt-link-rfc2396E"
                href="mailto:rpd@afrinic.net" moz-do-not-send="true">
                <rpd@afrinic.net></a>; <a
                class="x_moz-txt-link-abbreviated moz-txt-link-freetext"
                href="mailto:pdwg-chairs@afrinic.net"
                moz-do-not-send="true">
                pdwg-chairs@afrinic.net</a> <a
                class="x_moz-txt-link-rfc2396E"
                href="mailto:pdwg-chairs@afrinic.net"
                moz-do-not-send="true">
                <pdwg-chairs@afrinic.net></a><br>
              <b>Subject:</b> Re: [Last Call] Draft Policy Proposal -
              Hierarchical Names for New AS-SETs
              (AFPUB-2026-ASN-001-DRAFT02)</font>
            <div> </div>
          </div>
          <style type="text/css" style="display:none">p
        {margin-top:0;
        margin-bottom:0}</style>
          <div dir="ltr">
            <div class="x_x_elementToProof"
style="margin-top:0px; margin-bottom:0px; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              Dear Tshepo,</div>
            <div class="x_x_elementToProof"
style="margin-top:0px; margin-bottom:0px; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              <br>
            </div>
            <div class="x_x_elementToProof"
style="margin-top:0px; margin-bottom:0px; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              After careful consideration I have decided not to revise
              the draft. There are a couple of reasons for this:</div>
            <div class="x_x_elementToProof"
style="margin-top:0px; margin-bottom:0px; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              <br>
            </div>
            <ul
data-editing-info="{"applyListStyleFromLevel":false,"unorderedStyleType":1}"
style="margin-top:0px; margin-bottom:0px; list-style-type:disc">
              <li
style="font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0); margin-top:0px; margin-bottom:0px">
                <div class="x_x_elementToProof" role="presentation"
                  style="margin-top:0px; margin-bottom:0px">
                  I think this is becoming too prescriptive. If we try
                  to define a time period for restoration for example,
                  for some people it will be too long, for others too
                  short. We have no data to make an informed decision
                  here. Similarly the point about having a public log.
                  That requires us to define what that looks like and
                  how it works. We have no idea how often this process
                  would be enacted so we can't described what that needs
                  to look like. We have no data. I think trying to make
                  a perfect policy that covers all cases is the wrong
                  approach. I think we should be aiming to get
                  "something" implemented, let it run for a while,
                  gather feedback, and then if it needs adjusting, we
                  can raise another policy change proposal. Nothing is
                  set in stone here. There's no need to get his perfect
                  first time, we can iterate over time.</div>
                <div class="x_x_elementToProof" role="presentation"
                  style="margin-top:0px; margin-bottom:0px">
                  <br>
                </div>
              </li>
              <li
style="font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0); margin-top:0px; margin-bottom:0px">
                <div class="x_x_elementToProof" role="presentation"
                  style="margin-top:0px; margin-bottom:0px">
                  Any changes to the draft will restart the Last Call
                  and potentially also require the staff to re-review
                  the draft. The changes you proposed, whilst valid, are
                  optimising for a corner case (how often to people
                  delete their AS-SET by mistake? I would say, rarely).
                  Why delay getting "something" deployed, which we can
                  continue to iterate on, just to optimise for a corner
                  case (restores aren't blocked by this policy)?</div>
              </li>
            </ul>
            <div
style="margin-top:0px; margin-bottom:0px; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              <br>
            </div>
            <div class="x_x_elementToProof"
style="margin-top:0px; margin-bottom:0px; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              For these reasons I have decided not to modify the draft.</div>
            <div class="x_x_elementToProof"
style="margin-top:0px; margin-bottom:0px; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              <br>
            </div>
            <div class="x_x_elementToProof"
style="margin-top:0px; margin-bottom:0px; font-family:Calibri,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              With kind regards,</div>
            <div id="x_x_Signature" class="x_x_elementToProof">
              <div class="x_x_elementToProof"
style="font-family:Calibri,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
                James Bensley (he/him)</div>
            </div>
            <div
style="font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              <br>
            </div>
            <hr style="display:inline-block; width:98%">
            <div id="x_x_divRplyFwdMsg">
              <div
style="direction:ltr; font-family:Calibri,sans-serif; font-size:11pt; color:rgb(0,0,0)">
                <b>From:</b> Tshepo Masuku <a
                  class="x_moz-txt-link-rfc2396E"
                  href="mailto:TshepoMasuku26@hotmail.com"
                  moz-do-not-send="true">
                  <TshepoMasuku26@hotmail.com></a><br>
                <b>Sent:</b> 28 July 2026 15:28<br>
                <b>To:</b> James Bensley <a
                  class="x_moz-txt-link-rfc2396E"
                  href="mailto:james@inter.link" moz-do-not-send="true">
                  <james@inter.link></a>; <a
class="x_moz-txt-link-abbreviated moz-txt-link-freetext"
                  href="mailto:rpd@afrinic.net" moz-do-not-send="true">
                  rpd@afrinic.net</a> <a
                  class="x_moz-txt-link-rfc2396E"
                  href="mailto:rpd@afrinic.net" moz-do-not-send="true">
                  <rpd@afrinic.net></a>; <a
class="x_moz-txt-link-abbreviated moz-txt-link-freetext"
                  href="mailto:pdwg-chairs@afrinic.net"
                  moz-do-not-send="true">
                  pdwg-chairs@afrinic.net</a> <a
                  class="x_moz-txt-link-rfc2396E"
                  href="mailto:pdwg-chairs@afrinic.net"
                  moz-do-not-send="true">
                  <pdwg-chairs@afrinic.net></a><br>
                <b>Subject:</b> Re: [Last Call] Draft Policy Proposal -
                Hierarchical Names for New AS-SETs
                (AFPUB-2026-ASN-001-DRAFT02)</div>
              <div style="direction:ltr"> </div>
            </div>
            <p style="margin-top:0px; margin-bottom:0px"><span
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">Dear
                James,</span></p>
            <p style="margin-top:0px; margin-bottom:0px"><span
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">Thank
                you for engaging with the concern constructively. Your
                proposed wording is a meaningful improvement because it
                begins replacing an open-ended exception with objective
                conditions.</span></p>
            <p style="margin-top:0px; margin-bottom:0px"><span
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">I
                think a few points still need tightening before the rule
                is ready:</span></p>
            <ul>
              <li
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
                Restoration should return the object to its last valid
                archived state, not merely reuse the same key.
                Otherwise, an old flat name could be restored with
                entirely different contents and become a new object in
                substance.</li>
              <li
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
                The text should define whether the restoration right is
                permanent or subject to a recovery period.</li>
              <li
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
                “Another maintainer within the same organisation” and an
                organisation “responsible for the assets” require
                objective evidence, such as registry history or
                documented legal succession. Those terms should not be
                left entirely to staff interpretation.</li>
              <li
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
                The residual case-by-case provision should expressly
                prohibit creating a flat name that did not exist before
                implementation.</li>
              <li
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
                Every exceptional decision should record the evidence
                relied upon and provide a review or appeal path, not
                merely an audit entry.</li>
            </ul>
            <p style="margin-top:0px; margin-bottom:0px"><span
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">I
                would also suggest wording along these lines:</span></p>
            <blockquote>
              <p style="margin-top:0px; margin-bottom:0px"><span
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">A
                  restored object shall initially reflect its last valid
                  archived state, except that maintainer and contact
                  attributes may be updated where necessary to establish
                  control by the verified original holder or lawful
                  successor.</span></p>
            </blockquote>
            <p style="margin-top:0px; margin-bottom:0px"><span
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">Because
                this changes the normative enforcement boundary, I
                believe v3 should be published with an updated Impact
                Assessment and receive proper review. It should not be
                treated as a minor editorial correction during the final
                days of Last Call.</span></p>
            <p style="margin-top:0px; margin-bottom:0px"><span
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">I
                remain opposed to DRAFT02 as written, but I appreciate
                your willingness to address the defect directly. This is
                the kind of concrete engagement that adds value to the
                process.</span></p>
            <p style="margin-top:0px; margin-bottom:0px"><span
style="font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">Regards,<br>
                Tshepo</span></p>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              <br>
            </div>
            <div id="x_x_x_ms-outlook-mobile-body-separator-line">
              <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt">
                <br>
              </div>
            </div>
            <div id="x_x_x_ms-outlook-mobile-signature">
              <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
                <br>
              </div>
            </div>
            <hr style="display:inline-block; width:98%">
            <div id="x_x_x_divRplyFwdMsg">
              <div
style="direction:ltr; font-family:Calibri,sans-serif; font-size:11pt; color:rgb(0,0,0)">
                <b>From:</b> James Bensley <a
                  class="x_moz-txt-link-rfc2396E"
                  href="mailto:james@inter.link" moz-do-not-send="true">
                  <james@inter.link></a><br>
                <b>Sent:</b> Tuesday, 28 July 2026 13:39:50<br>
                <b>To:</b> Tshepo Masuku <a
                  class="x_moz-txt-link-rfc2396E"
                  href="mailto:TshepoMasuku26@hotmail.com"
                  moz-do-not-send="true">
                  <TshepoMasuku26@hotmail.com></a>; <a
class="x_moz-txt-link-abbreviated moz-txt-link-freetext"
                  href="mailto:rpd@afrinic.net" moz-do-not-send="true">
                  rpd@afrinic.net</a> <a
                  class="x_moz-txt-link-rfc2396E"
                  href="mailto:rpd@afrinic.net" moz-do-not-send="true">
                  <rpd@afrinic.net></a>; <a
class="x_moz-txt-link-abbreviated moz-txt-link-freetext"
                  href="mailto:pdwg-chairs@afrinic.net"
                  moz-do-not-send="true">
                  pdwg-chairs@afrinic.net</a> <a
                  class="x_moz-txt-link-rfc2396E"
                  href="mailto:pdwg-chairs@afrinic.net"
                  moz-do-not-send="true">
                  <pdwg-chairs@afrinic.net></a><br>
                <b>Subject:</b> Re: [Last Call] Draft Policy Proposal -
                Hierarchical Names for New AS-SETs
                (AFPUB-2026-ASN-001-DRAFT02)</div>
              <div style="direction:ltr"> </div>
            </div>
            <div style="direction:ltr">Dear <span
style="font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
                Tshepo,</span></div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              Thank you for providing your feedback. I am in agreement
              with your statement.</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              I am the original author of this draft change, and I wrote
              in my last update (version 2):</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              "Exceptions MUST be allowed on a case-by-case basis. For
              example, a non-hierarchically named AS-SET was deleted by
              mistake, so it should be possible to restore this AS-SET
              without having to rename it."</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              The staff assessment was to go with the following "7.8.7
              Exceptions for the creation of hierarchical names may be
              granted where necessary. AFRINIC shall document the reason
              for any exception.".</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              The staff asked me to review their assessment and I
              initially agreed, because it was loser than my definition.
              There may be other scenarios where restoration is required
              which I/we haven't thought of, so why limit it to
              accidental deletion only. Opening it up to allow any case
              to be reviewed has benefits. But as you say, it could also
              go the other way, it could simply mean that every case is
              rejected.</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              I am going to submit another version (v3) of the document,
              only replacing the text I quoted above, with the text
              below, to fix the problem you raised of the text having
              gone from being too strict before to now being too lose (I
              will try to propose something in the middle which
              guarantees delete/restoration is protected, anything else
              needs a review).</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              I am proposing the following text, what do you think?</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              ------</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              A non-hierarchical AS-SET existing before implementation
              of this policy change, which has since been deleted, must
              be restorable when the following conditions are met:</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              - The same object key is requested for restoration (no
              AS-SET name changes allowed)</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              - No object with a conflicting name has been created
              between the date of deletion and the date the restore is
              made</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              - The restore request is coming from either a maintainer
              that was present on the deleted object, or another
              maintainer within the same organisation (in the case the
              maintainer was also deleted), or a maintainer in an
              organisation responsible for the assets of the original
              organisation (in the case of a merger or acquisition).</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              - The restoration is recorded in an auditable log.</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              Any other request for restoring a deleted AS-SET created
              before the implementation date of this policy change will
              need to be reviewed by AFRINIC on a case by case basis.</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
              -------</div>
            <div id="x_x_x_x_Signature">
              <div
style="direction:ltr; font-family:Aptos,Aptos_EmbeddedFont,Aptos_MSFontService,Calibri,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
                <br>
              </div>
              <div
style="direction:ltr; font-family:Calibri,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
                With kind regards,</div>
              <div
style="direction:ltr; font-family:Calibri,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(0,0,0)">
                James Bensley (he/him)</div>
            </div>
            <hr style="direction:ltr; display:inline-block; width:98%">
            <div id="x_x_x_x_divRplyFwdMsg">
              <div
style="direction:ltr; font-family:Calibri,sans-serif; font-size:11pt; color:rgb(0,0,0)">
                <b>From:</b> Tshepo Masuku <a
                  class="x_moz-txt-link-rfc2396E"
                  href="mailto:TshepoMasuku26@hotmail.com"
                  moz-do-not-send="true">
                  <TshepoMasuku26@hotmail.com></a><br>
                <b>Sent:</b> 27 July 2026 15:29<br>
                <b>To:</b> <a
class="x_moz-txt-link-abbreviated moz-txt-link-freetext"
                  href="mailto:hvisage@hevis.co.za"
                  moz-do-not-send="true">
                  hvisage@hevis.co.za</a> <a
                  class="x_moz-txt-link-rfc2396E"
                  href="mailto:hvisage@hevis.co.za"
                  moz-do-not-send="true">
                  <hvisage@hevis.co.za></a>; Nonjabulo Sphilile <a
                  class="x_moz-txt-link-rfc2396E"
                  href="mailto:nonjabulosphilile@gmail.com"
                  moz-do-not-send="true">
                  <nonjabulosphilile@gmail.com></a>; Phetulo
                Dhlamini <a class="x_moz-txt-link-rfc2396E"
                  href="mailto:phetulodhlamini@gmail.com"
                  moz-do-not-send="true">
                  <phetulodhlamini@gmail.com></a>; Taye Medoye <a
                  class="x_moz-txt-link-rfc2396E"
                  href="mailto:daniel.medoye@gmail.com"
                  moz-do-not-send="true">
                  <daniel.medoye@gmail.com></a>; Thulisile Mazomba
                <a class="x_moz-txt-link-rfc2396E"
                  href="mailto:thulisile.mazomba@gmail.com"
                  moz-do-not-send="true">
                  <thulisile.mazomba@gmail.com></a>; <a
class="x_moz-txt-link-abbreviated moz-txt-link-freetext"
                  href="mailto:rpd@afrinic.net" moz-do-not-send="true">
                  rpd@afrinic.net</a> <a
                  class="x_moz-txt-link-rfc2396E"
                  href="mailto:rpd@afrinic.net" moz-do-not-send="true">
                  <rpd@afrinic.net></a><br>
                <b>Subject:</b> Re: [rpd] [Last Call] Draft Policy
                Proposal - Hierarchical Names for New AS-SETs
                (AFPUB-2026-ASN-001-DRAFT02)</div>
              <div style="direction:ltr"> </div>
            </div>
            <div
style="direction:ltr; background-color:rgb(255,243,205); margin-bottom:10px; padding:12px; border-width:1px; border-style:solid; border-color:rgb(255,238,186); font-family:Arial,sans-serif; font-size:14px; color:rgb(133,100,4)">
              <b>⚠️ Caution:</b> This email originated from outside of
              your organization. Do not click on links or open
              attachments unless you recognize the sender and know the
              content is safe.</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              Dear Hendrik and colleagues,</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              Thank you for consolidating the discussion. I will not
              repeat the earlier lifecycle, resolver, delegation, or
              membership arguments.</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              There is, however, a separate defect in the actual text
              being considered.</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              The proposal says that exceptions MUST be allowed on a
              case-by-case basis and gives restoration of an
              accidentally deleted flat AS-SET as the example. The
              staff-recommended wording instead says exceptions may be
              granted “where necessary.” These are materially different
              rules. </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              Under the first version, an eligible operator has a right
              to restoration. Under the second, restoration depends on
              AFRINIC’s discretion. The wording also changes a narrow
              restoration example into an undefined general exception.</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              That produces a concrete operational difference. If a
              grandfathered flat AS-SET is accidentally deleted after
              implementation, one version requires a restoration path
              while the other allows AFRINIC to refuse it. Operators and
              peers referencing that object cannot know which outcome
              the adopted policy guarantees.</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              This is not an adjacent lifecycle problem. It is the
              enforcement boundary of the present proposal.</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              The text should therefore be amended to provide one
              deterministic rule. For example:</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              > A non-hierarchical AS-SET existing before
              commencement may be restored only where the same object
              key is requested within a defined recovery period, the
              request is authenticated by the maintainer authorised
              immediately before deletion, no conflicting object has
              intervened, and the restoration is recorded in an
              auditable log. No other exception may authorise creation
              of a new non-hierarchical AS-SET.</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              That would preserve accidental-deletion recovery without
              leaving the supposedly closed namespace subject to
              undefined staff judgment.</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              Documenting the reason after an exception is granted is
              not the same as defining the conditions under which the
              exception is valid. Implementation guidance also cannot
              resolve a conflict between “MUST” and “may” after
              ratification.</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              Before Last Call closes, the authors and co-chairs should
              identify which wording is actually under consideration and
              resolve the difference in normative force.</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              Until then, I remain opposed to the proposal as written.</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              <br>
            </div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              Regards,</div>
            <div
style="direction:ltr; font-family:Aptos,Aptos_MSFontService,-apple-system,Roboto,Arial,Helvetica,sans-serif; font-size:12pt; color:rgb(33,33,33)">
              Tshepo </div>
            <hr style="direction:ltr; display:inline-block; width:98%">
            <div id="x_x_x_x_x_divRplyFwdMsg">
              <div
style="direction:ltr; font-family:Calibri,sans-serif; font-size:11pt; color:rgb(0,0,0)">
                <b>From:</b> <a
class="x_moz-txt-link-abbreviated moz-txt-link-freetext"
                  href="mailto:hvisage@hevis.co.za"
                  moz-do-not-send="true">
                  hvisage@hevis.co.za</a> <a
                  class="x_moz-txt-link-rfc2396E"
                  href="mailto:hvisage@hevis.co.za"
                  moz-do-not-send="true">
                  <hvisage@hevis.co.za></a><br>
                <b>Sent:</b> Monday, 27 July 2026 15:11:31<br>
                <b>To:</b> Tshepo Masuku <a
                  class="x_moz-txt-link-rfc2396E"
                  href="mailto:TshepoMasuku26@hotmail.com"
                  moz-do-not-send="true">
                  <TshepoMasuku26@hotmail.com></a>; Nonjabulo
                Sphilile <a class="x_moz-txt-link-rfc2396E"
                  href="mailto:nonjabulosphilile@gmail.com"
                  moz-do-not-send="true">
                  <nonjabulosphilile@gmail.com></a>; Phetulo
                Dhlamini <a class="x_moz-txt-link-rfc2396E"
                  href="mailto:phetulodhlamini@gmail.com"
                  moz-do-not-send="true">
                  <phetulodhlamini@gmail.com></a>; Taye Medoye <a
                  class="x_moz-txt-link-rfc2396E"
                  href="mailto:daniel.medoye@gmail.com"
                  moz-do-not-send="true">
                  <daniel.medoye@gmail.com></a>; Thulisile Mazomba
                <a class="x_moz-txt-link-rfc2396E"
                  href="mailto:thulisile.mazomba@gmail.com"
                  moz-do-not-send="true">
                  <thulisile.mazomba@gmail.com></a>; <a
class="x_moz-txt-link-abbreviated moz-txt-link-freetext"
                  href="mailto:rpd@afrinic.net" moz-do-not-send="true">
                  rpd@afrinic.net</a> <a
                  class="x_moz-txt-link-rfc2396E"
                  href="mailto:rpd@afrinic.net" moz-do-not-send="true">
                  <rpd@afrinic.net></a><br>
                <b>Subject:</b> [Last Call] Draft Policy Proposal -
                Hierarchical Names for New AS-SETs
                (AFPUB-2026-ASN-001-DRAFT02)</div>
              <div style="direction:ltr"> </div>
            </div>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">Good day Tshepo,
                Nonjabulo, Phetulo, Taye, Thulisile, colleagues,</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">Apologies to the human
                readers for this voluminous post, but I’d<br>
                rather address as much in one email than to flood the
                mailing list<br>
                With even more messages as I’d rather do a digest type
                response for<br>
                all the similar objections at once.</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">For nearly two weeks we
                have been asking for technical substance;<br>
                today, in the week Last Call closes, technical reasons
                have at last<br>
                been presented. Before I respond to them, let me add a
                seventh<br>
                question to the six of 23 July (015344):</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">Question 7: could you
                please provide the technical issues,<br>
                difficulties, and real-world examples where
                implementation of this<br>
                draft policy - as three RIRs have already implemented
                it, and the<br>
                fourth is formally proposing - would negatively affect
                your current<br>
                or future operations? We would like to understand those
                too.<br>
                Operational evidence of that kind would genuinely add to
                the body<br>
                of knowledge on this proposal.</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">Note: numbers like
                (015344) are message IDs in the RPD archive -<br>
                prepend </span><span
                style="font-family:sans-serif; color:rgb(57,131,196)"><a
                  href="https://lists.afrinic.net/pipermail/rpd/2026/"
                  id="OWAe288f940-a422-34b1-b13b-62d89dd1e2e2"
class="x_x_OWAAutoLink x_moz-txt-link-freetext moz-txt-link-freetext"
                  data-auth="NotApplicable"
style="color:rgb(57,131,196); margin-top:0px; margin-bottom:0px"
                  moz-do-not-send="true">https://lists.afrinic.net/pipermail/rpd/2026/</a></span><span
                style="font-family:sans-serif"> and append<br>
                .html to read any of them.</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">So let's get to the
                objections we have been waiting two weeks for:</span></p>
            <ol start="1" style="direction:ltr">
              <li style="font-family:sans-serif; direction:ltr">object
                lifecycle under changes of ASN control (015387,<br>
                escalated in 015391 and 015400) - answered by Jaco in
                (015388)<br>
                and (015394)</li>
            </ol>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">1b) its
                delegated-maintainer extension (015398, pressed again in<br>
                015400) - Jaco deferred to the RFC in (015399); the
                deferral is<br>
                completed in section 1 below<br>
                2) recursive expansion of members: references (015389) -
                answered<br>
                in (015390)<br>
                3) grandfathered flat objects (015392) - answered in
                (015395)<br>
                4) multi-source name resolution (015393) - answered in
                (015396,<br>
                015397)<br>
                5) anchor-ASN eligibility and authentication (015401) -
                addressed<br>
                in section 5 below<br>
                6) implementation rollout process (015402) - addressed
                in section 6<br>
                below; its delegation and 7.8.7 points are answered in
                section 1<br>
                (the delegation question is 015398/015400 returning),
                and its<br>
                benefit-scope point in the common ground just below</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">I refer to and endorse
                Jaco's answers rather than repeat them. It<br>
                was asked earlier in this Last Call that objections be
                addressed as<br>
                grouped issues rather than piecemeal, and in that spirit
                I have<br>
                collated the responses in a single email, adding what
                has not yet<br>
                been shared: a precise statement of the authorisation
                model, the<br>
                published data, and the cross-registry record.</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">First, the common ground,
                because it is substantial. In today's own<br>
                words:</span></p>
            <ul style="direction:ltr">
              <li style="font-family:sans-serif; direction:ltr">hierarchical
                naming "can strengthen creation authorisation" (015393)</li>
              <li style="font-family:sans-serif; direction:ltr">the
                proposal "secures the first object name" (015389)</li>
              <li style="font-family:sans-serif; direction:ltr">the
                concerns are "not questions about whether AS-SET
                membership<br>
                is correct" (015387)</li>
              <li style="font-family:sans-serif; direction:ltr">ASN-anchored
                naming "can reduce future flat-name collisions"<br>
                (015402)</li>
            </ul>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">None of today's messages
                disputes what the proposal does. Each of<br>
                today's objections asks it to also solve an adjacent
                problem -</span></p>
            <ul style="direction:ltr">
              <li style="font-family:sans-serif; direction:ltr">lifecycle</li>
              <li style="font-family:sans-serif; direction:ltr">delegation
                semantics</li>
              <li style="font-family:sans-serif; direction:ltr">expansion
                semantics</li>
              <li style="font-family:sans-serif; direction:ltr">legacy
                objects</li>
              <li style="font-family:sans-serif; direction:ltr">mirror
                copies</li>
              <li style="font-family:sans-serif; direction:ltr">anchor
                eligibility</li>
              <li style="font-family:sans-serif; direction:ltr">and
                every one of those problems exists today, everywhere,
                with or</li>
            </ul>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">without this policy. That
                is not the discovery of a defect. It is<br>
                the discovery that the proposal has a scope, which it
                states and<br>
                which was put on this record weeks ago (015119, 015120,
                015123; and<br>
                the draft's own sections 7.8.3-7.8.6).</span></p>
            <ol start="1" style="direction:ltr">
              <li style="font-family:sans-serif; direction:ltr">The
                authorisation model, changes of ASN control, and
                delegation<br>
                (015387, 015391, 015398, 015400, 015402)</li>
            </ol>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">Jaco deferred to the RFC
                on the depth mechanics (015399). Rightly<br>
                so - the RFC, followed to the top of the chain,
                completes his<br>
                answer. (015398) states the rule correctly: a
                first-level object,<br>
                AS12345:AS-CUSTOMERS, is created under the authorisation
                of the<br>
                aut-num AS12345 as it stands at that moment - its
                maintainer chain,<br>
                which follows the current registered holder. A deeper
                object,<br>
                AS12345:AS-CUSTOMERS:AS-EU, is created under the
                authorisation of<br>
                the maintainer of its immediate parent. Now ask where
                that parent<br>
                came from: its own creation was authorised by the level
                above it,<br>
                and so on, terminating at the aut-num itself. Every
                object in the<br>
                tree exists because the level above authorised its
                creation.<br>
                Delegation, where it exists, exists because the chain
                authorised<br>
                it. And who controls each level is answerable, level by
                level, from<br>
                the database as it stands - which is what verifiable
                control looks<br>
                like in RPSL.</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">The emphatic difference
                at the current state is that a flat name<br>
                offers just a maintainer but no chain: no root, and no
                way to ask<br>
                whether the name was ever anyone's to create - beyond
                that it must<br>
                not already exist in THIS one database.</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">(015400) concludes that
                because deeper control can be delegated,<br>
                the "control model remains unspecified". The opposite is
                the case:<br>
                the control model is fully specified - in the very RFC
                that<br>
                (015398) cited, which is why deferring to it was the
                right answer.<br>
                That deeper authority is delegated, and that delegated
                authority<br>
                does not automatically follow the ASN, is not a gap in
                the model;<br>
                it is the model, working as designed for twenty-seven
                years, for<br>
                every hierarchical object family, at every RPSL-based
                registry. A<br>
                policy does not "leave AFRINIC to interpret" what the
                RFC and the<br>
                production database already define - any more than this
                proposal<br>
                needs to restate what mnt-by means. What
                deterministically follows<br>
                the ASN is the root of every chain: authority over
                creation and<br>
                recreation at the first level of the namespace. That is
                precisely</span></p>
            <ul style="direction:ltr">
              <li style="font-family:sans-serif; direction:ltr">and only
                - what the proposal's attribution claim covers. The same</li>
            </ul>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">depth question returns in
                (015402); the answer above stands.</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">These semantics are not
                invented by this proposal. They are RFC<br>
                2622 (1999), operated by RPSL-based registries -
                AFRINIC's included</span></p>
            <ul style="direction:ltr">
              <li style="font-family:sans-serif; direction:ltr">for
                twenty-seven years. The proposal does not modify them;
                it</li>
            </ul>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">gates which names may
                come into existence, nothing else. The two<br>
                outcomes (015398) poses are therefore both answered by
                the<br>
                semantics already in production: descendants do not move<br>
                automatically, so no delegated maintainer is overridden
                - exactly<br>
                as for every delegated object today. And the new holder
                gains<br>
                precisely what the name anchors: authority over creation
                and<br>
                recreation at the first level of its namespace. What the
                new holder<br>
                does not get - control over objects a previous holder's
                delegates<br>
                still maintain - is what nobody has today under flat
                names, for any<br>
                object, ever. A stale delegated descendant is today's
                universal<br>
                condition; the proposal neither creates nor worsens it,
                and no<br>
                registry in those twenty-seven years has defined the
                demanded<br>
                delegation-and-revocation regime for set objects in
                policy.</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">How ASN control actually
                changes in this region is likewise not a<br>
                matter for speculation. AFRINIC publishes it:</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif; color:rgb(57,131,196)"><a
href="https://ftp.afrinic.net/pub/stats/afrinic/transfers/transfers_latest.json"
                  id="OWAe08156bb-9a61-a99c-74ff-99749c6c5545"
class="x_x_OWAAutoLink x_moz-txt-link-freetext moz-txt-link-freetext"
                  data-auth="NotApplicable"
style="color:rgb(57,131,196); margin-top:0px; margin-bottom:0px"
                  moz-do-not-send="true">https://ftp.afrinic.net/pub/stats/afrinic/transfers/transfers_latest.json</a></span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">A full read of that file
                today: 75 ASN-carrying transfer events (91<br>
                ASNs) from 2018 through 2026 - about nine such events a
                year - and<br>
                every single one is a merger/acquisition succession, in
                which the<br>
                organisation and its maintainer control pass together,
                so parent<br>
                ASN and child objects move as one unit. That is the
                published<br>
                record behind Jaco's description in (015394), with the
                frequency<br>
                question answered exactly: none open-market, all
                successions that<br>
                preserve the attribution chain by construction. Anyone
                can verify<br>
                this from the file. So the concern of (015391) has been
                addressed -<br>
                with our registry's own data.</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">The cross-registry record
                answers the rest. I checked this week,<br>
                across the policy, procedural and database-documentation
                layers of<br>
                the other four RIRs. RIPE has required hierarchical
                names for new<br>
                as-sets since December 2022 and operates full ASN
                transfers,<br>
                including inter-RIR; APNIC likewise since July 2023
                (prop-151);<br>
                LACNIC's IRR has been structurally hierarchical since it
                launched<br>
                in 2020; at ARIN hierarchical naming is available today
                and a<br>
                mandate for new objects is formally proposed
                (ARIN-prop-342,<br>
                pending) - the same direction. All four are silent on
                child as-set<br>
                lifecycle at every layer. Across the deployed
                implementations that<br>
                is roughly three and a half years of exactly the
                coexistence<br>
                (015387) warns about - hierarchical mandates running
                over live ASN<br>
                transfer markets - and I could not find one documented
                incident of<br>
                the failure mode described. In every deployed
                implementation,<br>
                lifecycle lives where it belongs: in standard RPSL
                authorisation<br>
                and registry operations, not in naming-policy text.</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">On 7.8.7 (015391,
                015402): today the namespace has no rule at all -<br>
                every creation, by anyone, of any name, is the
                exception. A closed<br>
                default with documented-reason exceptions cannot be less<br>
                deterministic than no default whatsoever. And credit
                where due:<br>
                (015402) makes one point that verifies - the staff
                assessment's<br>
                prose cites "7.8.6" for the
                restoration-of-deleted-objects<br>
                exception, where 7.8.6 concerns editing and the
                exception clause is<br>
                in fact 7.8.7. That is a cross-reference slip in the
                assessment's<br>
                commentary, worth an erratum, and it changes nothing in
                the policy<br>
                text: the clauses say what they say, and staff's own
                recorded<br>
                caution about recall loopholes shows the right mechanism
                was<br>
                analysed. As for the demanded exception criteria: 7.8.7
                requires<br>
                every exception to be documented with its reason - a
                discipline<br>
                stricter than today, where no creation needs any reason
                at all.</span></p>
            <ol start="2" style="direction:ltr">
              <li style="font-family:sans-serif; direction:ltr">Recursive
                expansion of members: references (015389)</li>
            </ol>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">Correct as far as it goes
                - and it concedes the top level is<br>
                secured. What members: may contain is a content
                question, expressly<br>
                outside this proposal's scope, and outside the naming
                policies of<br>
                every registry that already runs this rule: none coupled
                naming to<br>
                expansion semantics, because that coupling is a
                different and<br>
                larger policy. Jaco has already named the constructive
                path<br>
                (015390): a members-qualification rule would be a
                coherent separate<br>
                proposal, and wording was invited. Meanwhile every new
                hierarchical<br>
                set is one fewer ambiguous reference for the next
                expansion to fear</span></p>
            <ul style="direction:ltr">
              <li style="font-family:sans-serif; direction:ltr">the gap
                this concern describes narrows with adoption and stays</li>
            </ul>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">open without it.</span></p>
            <ol start="3" style="direction:ltr">
              <li style="font-family:sans-serif; direction:ltr">Grandfathered
                flat objects and the transition window (015392)</li>
            </ol>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">The capability described
                - register flat names now, populate them<br>
                later - exists today, permanently, with or without this
                proposal.<br>
                Anyone may do it this afternoon. The proposal is the
                only mechanism<br>
                before this Working Group that ever closes that window;
                declining<br>
                it preserves the window forever. If gaming of the
                implementation<br>
                gap is the genuine worry, the remedy is a short
                implementation<br>
                period - a staff scheduling matter - not an indefinitely
                open<br>
                namespace. And, respectfully: (015392) describes with
                precision the<br>
                squatting behaviour whose future possibility this policy
                exists to<br>
                end. An objection that demonstrates the attack is an
                argument for<br>
                closing the door.</span></p>
            <ol start="4" style="direction:ltr">
              <li style="font-family:sans-serif; direction:ltr">Multi-source
                resolution (015393)</li>
            </ol>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">Jaco's answer (015396,
                015397) is complete: an ASN is delegated by<br>
                exactly one registry, so a hierarchical name carries its<br>
                authoritative source in its own first element. The
                question "which<br>
                source is authoritative for this set?" has a
                deterministic answer<br>
                precisely and only when the name is hierarchical; under
                a flat name<br>
                that answer does not exist anywhere. The working-group
                draft cited<br>
                in (015393) pursues the same determinism through
                source-qualified<br>
                references - the two mechanisms are complements, not
                conflicts, and<br>
                that draft's problem statement describes the flat-name
                condition<br>
                this proposal removes for new AFRINIC sets. What a
                third-party<br>
                mirror carries is today's condition for every object in
                every IRR<br>
                database and is governed by no naming policy at any
                registry.</span></p>
            <ol start="5" style="direction:ltr">
              <li style="font-family:sans-serif; direction:ltr">Anchor-ASN
                eligibility and authentication (015401)</li>
            </ol>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">Taye, this is the kind of
                message Last Call is for: specific,<br>
                technical, and naming the exact requirements it wants.
                Three<br>
                responses.</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">First, on substance: the
                properties you list - the leading ASN<br>
                globally assigned, an authoritative aut-num present,
                creation<br>
                authenticated against that object's maintainer,
                private-use and<br>
                reserved ranges (RFC 6996) excluded, canonical ASPLAIN
                form (RFC<br>
                5396) - are, I would argue, what the Impact Assessment's
                own<br>
                authorisation claim already entails. An anchor with no
                holder can<br>
                authorise no one; a check that can pass in the absence
                of the<br>
                aut-num it authenticates against would not be the check
                the<br>
                assessment describes. So I read your list not as a new
                mechanism<br>
                but as the strict, and correct, reading of the mechanism
                the<br>
                proposal already claims.</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">Second, on where it
                belongs: those are implementation semantics -<br>
                the authentication mode of the database software and the
                eligible<br>
                ASN ranges - and I would support AFRINIC staff recording
                exactly<br>
                your list in the implementation guidance for this
                policy, alongside<br>
                the failure behaviour per creation path. That is the
                normal home of<br>
                such requirements at every registry, and section 7.8.7's<br>
                documented-exception mechanism gives staff the
                instrument for any<br>
                edge the guidance must handle. It is worth adding: even
                in the<br>
                weakest imaginable configuration, an unattributable<br>
                hierarchical-form name is no worse than today, where
                every name of<br>
                any form requires no authorisation from anyone.</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">Third, on form: yours is
                the closest any message in this Last Call<br>
                has come to proposed text. If you formalise that list as
                wording -<br>
                whether as implementation guidance or as a clarification
                for a<br>
                future revision - I have no doubt this Working Group
                will engage it<br>
                on its merits, exactly as was invited days ago (015390).
                It is the<br>
                constructive instrument the process recognises.</span></p>
            <ol start="6" style="direction:ltr">
              <li style="font-family:sans-serif; direction:ltr">Implementation
                rollout process (015402)</li>
            </ol>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">Warning-only phases, test
                vectors, baselines, rollback plans and<br>
                scheduled reviews are implementation artifacts. They are
                produced<br>
                by registry staff during implementation - at every RIR -
                and they<br>
                appear in no naming policy anywhere: RIPE shipped its
                hierarchical<br>
                requirement through the database release process, with
                release<br>
                notes and a release-candidate test environment, on the
                strength of<br>
                a policy text that contains none of those artifacts.
                Nothing in<br>
                this proposal forbids a staged, warning-first
                deployment; that is<br>
                exactly the kind of detail the implementation phase
                exists to<br>
                define, and 7.8.7's documented-exception mechanism gives
                staff the<br>
                instrument for edge handling during it. Taken at face
                value,<br>
                (015402)'s final section is an argument about deployment<br>
                scheduling, which follows ratification - not a reason to
                withhold<br>
                it.</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">Stepping back, twice.</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">Each of the demands in
                messages (015387), (015389), (015391),<br>
                (015392), (015393), (015398), (015400) and (015402)
                carries the same<br>
                condition: "before adoption" (015387, 015389, 015391),
                "before<br>
                mandatory enforcement" (015393), "before this proceeds"
                (015398),<br>
                "until ... addressed" (015392), "until that is
                clarified" (015400),<br>
                "until ... clearly resolved" (015402).<br>
                As Seun reminded the list (015386), this PDP has no
                conditional<br>
                adoption: a proposal cannot pass Last Call subject to
                future<br>
                homework. At day eleven of Last Call, "resolve and test
                before<br>
                advancing" therefore spells "do not adopt". The process
                has exactly<br>
                one instrument for a genuine, fixable concern: proposed
                text. Jaco<br>
                invited it (015390).</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">Let me emphasise the
                arithmetic: seven distinct technical concerns<br>
                in a single day, after eleven days without one - and,
                section 5's requirements<br>
                list aside, still zero proposed amendments. Of the six
                questions of<br>
                23 July (015344), exactly one account engaged them point
                by point -<br>
                Tshepo, in (015348) - and the record should credit those
                answers.<br>
                In short:</span></p>
            <ol start="1" style="direction:ltr">
              <li style="font-family:sans-serif; direction:ltr">Two-object
                case: tooling should bind to an explicitly selected<br>
                source or stop and report - the fail-closed behaviour
                supporters<br>
                had proposed; the live case itself stays unresolved
                either way,<br>
                the rule being prospective.</li>
              <li style="font-family:sans-serif; direction:ltr">Squatting
                remedy: none - the victim is left to another<br>
                database's dispute or abuse process.</li>
              <li style="font-family:sans-serif; direction:ltr">Is it an
                improvement? "Yes, narrowly."</li>
              <li style="font-family:sans-serif; direction:ltr">Evidence
                bar: a four-part test - whose second criterion, that<br>
                the proposal "would have prevented" the incident,
                excludes by<br>
                construction every incident that already exists; a
                creation-time<br>
                rule is prospective by definition.</li>
              <li style="font-family:sans-serif; direction:ltr">Impact
                Assessment: "I do not dispute the findings"; the<br>
                objection shifts to what the IA does not cover.</li>
              <li style="font-family:sans-serif; direction:ltr">Capacity:
                individual, representing no organisation or resource<br>
                member.</li>
            </ol>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">No other opposing account
                has answered any of the six. Today the<br>
                bar has moved with nearly every message, and a seventh
                question now<br>
                joins the list. My assessment, for the record:<br>
                none of today's objections identifies a defect in what
                the proposal<br>
                claims; and their cumulative effect, whatever the
                intent, has been<br>
                to consume the finite attention of this Working Group -
                a cost<br>
                long-standing participants have already described in
                their own<br>
                words (015154, 015179).</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">One further observation,
                offered for the record and without<br>
                attributing it to any individual: this Last Call's
                objections began<br>
                from the principle that the registry "may record... It
                may not<br>
                rule" (015011). Today's objections ask the registry to
                pre-define<br>
                delegation graphs, revocation processes, state machines,
                permanent<br>
                reservations, transition audits and source-selection
                rules, in<br>
                advance (015391, 015392, 015393, 015398, 015400,
                015402). Those two<br>
                standards pull in opposite directions, and the same
                seven-clause<br>
                creation-time rule cannot be faulted under both at once.
                Which<br>
                standard the Working Group actually holds is exactly
                what the<br>
                Co-Chairs will weigh - against what the proposal
                actually claims: a<br>
                creation-time naming rule, whose every stated boundary
                today's<br>
                objection messages ask it to exceed.</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">The community is entitled
                to expect Last Call to close on schedule<br>
                and to be assessed on the record as it stands.</span></p>
            <p style="direction:ltr; margin-top:0px; margin-bottom:0px"><span
                style="font-family:sans-serif">Regards,<br>
                Hendrik Visage (AS329532; also AS213481)</span></p>
            <div style="direction:ltr; display:none">[CompanySignature]</div>
            <div
style="direction:ltr; font-family:"Calibri",sans-serif; font-size:8pt">Inter<span
                style="font-size:0px">.</span>.link GmbH
              <b>|</b> Boxhagener Straße 80, 10245 Berlin, Germany <b>|</b>
              Managing Directors: Marc Korthaus, Theo Voss
              <b>|</b> Commercial Register: Amtsgericht Charlottenburg,
              HRB 138876 <b>|</b> VAT ID: DE281288887
              <b>|</b> Email: <a href="mailto:hello@inter.link"
                id="OWA02ff23ec-3d1e-7759-6597-38c0b37631c4"
class="x_x_OWAAutoLink x_moz-txt-link-freetext moz-txt-link-freetext"
                moz-do-not-send="true">
                hello@inter.link</a> <b>|</b> Web: <a
                href="https://inter.link"
                id="OWAf8d7f1f5-4be0-1f7a-4593-690df3987cb8"
                class="x_x_OWAAutoLink" data-auth="NotApplicable"
                moz-do-not-send="true">
                inter.link</a></div>
          </div>
          <br>
          <fieldset class="x_moz-mime-attachment-header"></fieldset>
          <pre class="x_moz-quote-pre">_______________________________________________
RPD mailing list
<a class="x_moz-txt-link-abbreviated moz-txt-link-freetext"
          href="mailto:RPD@afrinic.net" moz-do-not-send="true">RPD@afrinic.net</a>
<a class="x_moz-txt-link-freetext moz-txt-link-freetext"
          href="https://lists.afrinic.net/mailman/listinfo/rpd"
          moz-do-not-send="true">https://lists.afrinic.net/mailman/listinfo/rpd</a>
</pre>
        </blockquote>
      </div>
    </blockquote>
  </body>
</html>