<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    Not having these CAA records is a permission to issue. No DNS access
    is required to adjust anything.<br>
    <br>
    If a concerned Domain Name owner wants to use CAA to restrict
    issuance to a specific set of CAs, they better know what they're
    doing because it might disable their ability to get publicly trusted
    certificates.<br>
    <br>
    It's just like setting up SPF/DKIM (the mail system works fine
    without it). The mail admin should know what to do otherwise mails
    might be lost if something is not configured properly. <br>
    <br>
    <br>
    Dimitris.<br>
    <br>
    <div class="moz-cite-prefix">On 30/1/2021 9:23 π.μ., Paul van
      Brouwershaven wrote:<br>
    </div>
    <blockquote type="cite"
cite="mid:DM5PR11MB00738FDDCBEBE718F4D499A0F8B89@DM5PR11MB0073.namprd11.prod.outlook.com">
      <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
      <div dir="auto" style="direction: ltr; margin: 0; padding: 0;
        font-family: sans-serif; font-size: 11pt; color: black; ">
        Looking at it from a commercial point of view I get the interest
        to separate all type of certificates (everything that prevents
        issuance costs money). From a security point of view it's sounds
        like a firewall that only allows me to block known issues.<br>
        <br>
      </div>
      <div dir="auto" style="direction: ltr; margin: 0; padding: 0;
        font-family: sans-serif; font-size: 11pt; color: black; ">
        As in your firewall, you want to deny all and then start
        allowing what should pas through.
        <br>
        <br>
      </div>
      <div dir="auto" style="direction: ltr; margin: 0; padding: 0;
        font-family: sans-serif; font-size: 11pt; color: black; ">
        Allowing something new could be a internal request like you
        would for for all dns records (like you also do not give the
        whole company access to your dns system or your firewall and
        most corporate devices don't allow users to install any
        applications without prior approval).<br>
        <br>
      </div>
      <div dir="auto" style="direction: ltr; margin: 0; padding: 0;
        font-family: sans-serif; font-size: 11pt; color: black; ">
        If you let everything though you are out of control. Like in the
        early days of the internet we worked with block lists. Today
        everything is based on permit list, including SPF, DKIM and
        actually the CAA RFC.<br>
        <br>
      </div>
      <div dir="auto" style="direction: ltr; margin: 0; padding: 0;
        font-family: sans-serif; font-size: 11pt; color: black; ">
        Should we not focus on security and control above convenience
        and commercial interest?<br>
      </div>
      <div dir="auto" style="direction: ltr; margin: 0; padding: 0;
        font-family: sans-serif; font-size: 11pt; color: black; ">
        <span id="ms-outlook-android-cursor"></span><br>
      </div>
      <div dir="auto" style="direction: ltr; margin: 0; padding: 0;
        font-family: sans-serif; font-size: 11pt; color: black; ">
        <span id="OutlookSignature">
          <div dir="auto" style="direction: ltr; margin: 0; padding: 0;
            font-family: sans-serif; font-size: 11pt; color: black; ">
            Sent from my mobile device.</div>
        </span></div>
      <hr style="display:inline-block;width:98%" tabindex="-1">
      <div id="divRplyFwdMsg" dir="ltr"><font style="font-size:11pt"
          face="Calibri, sans-serif" color="#000000"><b>From:</b>
          Dimitris Zacharopoulos (HARICA) <a class="moz-txt-link-rfc2396E" href="mailto:dzacharo@harica.gr"><dzacharo@harica.gr></a><br>
          <b>Sent:</b> Saturday, January 30, 2021 7:57:34 AM<br>
          <b>To:</b> Paul van Brouwershaven
          <a class="moz-txt-link-rfc2396E" href="mailto:Paul.vanBrouwershaven@entrust.com"><Paul.vanBrouwershaven@entrust.com></a>; SMIME Certificate
          Working Group <a class="moz-txt-link-rfc2396E" href="mailto:smcwg-public@cabforum.org"><smcwg-public@cabforum.org></a>; Tim Hollebeek
          <a class="moz-txt-link-rfc2396E" href="mailto:tim.hollebeek@digicert.com"><tim.hollebeek@digicert.com></a>; Neil Dunbar
          <a class="moz-txt-link-rfc2396E" href="mailto:ndunbar@trustcorsystems.com"><ndunbar@trustcorsystems.com></a><br>
          <b>Cc:</b> Kirk Hall <a class="moz-txt-link-rfc2396E" href="mailto:Kirk.Hall@entrust.com"><Kirk.Hall@entrust.com></a><br>
          <b>Subject:</b> [EXTERNAL] Re: [Smcwg-public] CAA and S/MIME</font>
        <div> </div>
      </div>
      <div>
        <div>WARNING: This email originated outside of Entrust.<br>
          DO NOT CLICK links or attachments unless you trust the sender
          and know the content is safe.<br>
        </div>
        I also believe that Domain Name owners should have a different
        tag per certificate type. It doesn't make sense to have a "one
        tag to rule them all".<br>
        <br>
        So, if an owner wants to restrict ALL certificate issuance to
        one CA for TLS and S/MIME, two CAA records would need to be
        added.<br>
        <br>
        <br>
        Dimitris.<br>
        <br>
        <div class="x_moz-cite-prefix">On 29/1/2021 9:08 μ.μ., Paul van
          Brouwershaven via Smcwg-public wrote:<br>
        </div>
        <blockquote type="cite">
          <div dir="auto" style="color:rgb(33,33,33);
            background-color:rgb(255,255,255); text-align:left">
            <span style="font-size:12pt">Do I understand you correctly
              that in your opinion domain name owners should not have
              the ability to restrict all certificate issuance with a
              single record?</span><br>
          </div>
          <div dir="auto" style="color:rgb(33,33,33);
            background-color:rgb(255,255,255); text-align:left">
            <span style="font-size:12pt"><br>
            </span></div>
          <div dir="auto" style="color:rgb(33,33,33);
            background-color:rgb(255,255,255); text-align:left">
            I don't think we can expect that a domain name owner would
            add CAA records for every ecosystem (if they even know about
            there existence) because these ecosystems want to be
            independent.</div>
          <div dir="auto" style="color:rgb(33,33,33);
            background-color:rgb(255,255,255); text-align:left">
            <br>
          </div>
          <div dir="auto" style="color:rgb(33,33,33);
            background-color:rgb(255,255,255); text-align:left">
            If you have a link to the relevant discussion on MDSP that
            would be great!</div>
          <div dir="auto" style="color:rgb(33,33,33);
            background-color:rgb(255,255,255); text-align:left">
            <br>
          </div>
          <div id="x_id-c49b608e-59a5-482a-9a5b-0b4c88792ac7"
            class="x_ms-outlook-mobile-reference-message">
            <hr tabindex="-1" style="display:inline-block; width:98%">
            <div id="x_divRplyFwdMsg"><strong>From:</strong> Tim
              Hollebeek <a class="x_moz-txt-link-rfc2396E"
                href="mailto:tim.hollebeek@digicert.com"
                moz-do-not-send="true">
                <tim.hollebeek@digicert.com></a><br>
              <strong>Sent:</strong> Friday, 29 January 2021, 19:18<br>
              <strong>To:</strong> Paul van Brouwershaven; Neil Dunbar;
              SMIME Certificate Working Group<br>
              <strong>Cc:</strong> Kirk Hall<br>
              <strong>Subject:</strong> [EXTERNAL] RE: [Smcwg-public]
              CAA and S/MIME<br>
            </div>
            <br>
            <meta content="text/html; charset=Windows-1252">
            <meta name="Generator" content="Microsoft Word 15 (filtered
              medium)">
            <style>@font-face
        {font-family:Wingdings}@font-face
        {font-family:"Cambria Math"}@font-face
        {font-family:Calibri}@font-face
        {font-family:Consolas}@font-face
        {font-family:"Segoe UI"}p.x_MsoNormal, li.x_MsoNormal, div.x_MsoNormal
        {margin:0in;
        font-size:11.0pt;
        font-family:"Calibri",sans-serif}a:link, span.x_MsoHyperlink
        {color:blue;
        text-decoration:underline}pre
        {margin:0in;
        margin-bottom:.0001pt;
        font-size:10.0pt;
        font-family:"Courier New"}span.x_HTMLPreformattedChar
        {font-family:Consolas}p.x_xmsonormal, li.x_xmsonormal, div.x_xmsonormal
        {margin:0in;
        font-size:11.0pt;
        font-family:"Calibri",sans-serif}p.x_xmsolistparagraph, li.x_xmsolistparagraph, div.x_xmsolistparagraph
        {margin-right:0in;
        margin-left:0in;
        font-size:11.0pt;
        font-family:"Calibri",sans-serif}span.x_EmailStyle24
        {font-family:"Calibri",sans-serif;
        color:windowtext}.x_MsoChpDefault
        {font-size:10.0pt}ol
        {margin-bottom:0in}ul
        {margin-bottom:0in}</style>
            <div>WARNING: This email originated outside of Entrust.<br>
              DO NOT CLICK links or attachments unless you trust the
              sender and know the content is safe.<br>
            </div>
            <div class="x_WordSection1">
              <p class="x_MsoNormal">Paul,</p>
              <p class="x_MsoNormal"> </p>
              <p class="x_MsoNormal">You might want to review the
                previous discussions of this issue on MDSP, where it was
                made pretty clear, including by the author of the RFC,
                that the “issue” tag was intended to be specific to the
                web.  CAA for certificate type has also been discussed
                quite a bit here at the Forum recently (it was an idea
                we introduced about three years ago and were pushing),
                so you might want to review the long discussion of those
                proposals and why they didn’t move forward.</p>
              <p class="x_MsoNormal"> </p>
              <p class="x_MsoNormal">It’s not clear why you think there
                are problems with having both ‘issue’ and ‘issueesmime’,
                especially since your analysis seems to assume that if
                they’re both there, they interact in some way.  They
                should not, and that seems to be the source of the
                problems you’re trying to highlight.</p>
              <p class="x_MsoNormal"> </p>
              <p class="x_MsoNormal">What one wants is to be able to
                clearly state the policy for each ecosystem, without
                interactions.  Interactions between different
                certificate ecosystems are the cause of most of PKIs
                problems, and we should be looking to eliminate
                cross-PKI interactions, not introduce new ones.</p>
              <p class="x_MsoNormal"> </p>
              <p class="x_MsoNormal">It’s pretty straightforward to do
                that with a new tag.  No additional properties or
                semantics are required.  And the process of adding a new
                tag is something this group has already successfully
                done once (“CAA CONTACT”).</p>
              <p class="x_MsoNormal"> </p>
              <p class="x_MsoNormal">-Tim</p>
              <p class="x_MsoNormal"> </p>
              <div style="border:none; border-left:solid blue 1.5pt;
                padding:0in 0in 0in 4.0pt">
                <div>
                  <div style="border:none; border-top:solid #E1E1E1
                    1.0pt; padding:3.0pt 0in 0in 0in">
                    <p class="x_MsoNormal"><b>From:</b> Paul van
                      Brouwershaven <a class="x_moz-txt-link-rfc2396E"
                        href="mailto:Paul.vanBrouwershaven@entrust.com"
                        moz-do-not-send="true">
                        <Paul.vanBrouwershaven@entrust.com></a> <br>
                      <b>Sent:</b> Friday, January 29, 2021 11:13 AM<br>
                      <b>To:</b> Neil Dunbar <a
                        class="x_moz-txt-link-rfc2396E"
                        href="mailto:ndunbar@trustcorsystems.com"
                        moz-do-not-send="true">
                        <ndunbar@trustcorsystems.com></a>; SMIME
                      Certificate Working Group <a
                        class="x_moz-txt-link-rfc2396E"
                        href="mailto:smcwg-public@cabforum.org"
                        moz-do-not-send="true">
                        <smcwg-public@cabforum.org></a>; Tim
                      Hollebeek <a class="x_moz-txt-link-rfc2396E"
                        href="mailto:tim.hollebeek@digicert.com"
                        moz-do-not-send="true">
                        <tim.hollebeek@digicert.com></a><br>
                      <b>Cc:</b> Kirk Hall <a
                        class="x_moz-txt-link-rfc2396E"
                        href="mailto:Kirk.Hall@entrust.com"
                        moz-do-not-send="true">
                        <Kirk.Hall@entrust.com></a><br>
                      <b>Subject:</b> Re: [Smcwg-public] CAA and S/MIME</p>
                  </div>
                </div>
                <p class="x_MsoNormal"> </p>
                <div>
                  <p class="x_MsoNormal"><span style="font-size:12.0pt;
                      color:black">While the BR only specifies how CAA
                      must be implemented/used for TLS certificates the
                      CAA RFC is not limited to just TLS certificates,
                      the RFC 8659 (and previously RFC 6844) begins
                      with:</span></p>
                </div>
                <div>
                  <div>
                    <p class="x_MsoNormal" style="background:white"><span
                        style=""> </span></p>
                    <blockquote>
                      <pre style="background:white"><span style="color:black">The Certification Authority Authorization (CAA) DNS Resource Record</span></pre>
                      <pre style="background:white"><span style="color:black">   allows a DNS domain name holder to specify one or more Certification</span></pre>
                      <pre style="background:white"><span style="color:black">   Authorities (CAs) authorized to issue certificates for that domain</span></pre>
                      <pre style="background:white"><span style="color:black">   name.</span></pre>
                    </blockquote>
                    <p class="x_MsoNormal" style="background:white"><span
                        style=""> </span></p>
                  </div>
                  <p class="x_MsoNormal"><span style="font-size:12.0pt;
                      color:black">This does make sense as this would
                      create a `deny all, except` principle where you
                      need to give explicit permission like as in a good
                      firewall configuration. But I agree that there
                      should be a possibility to change permission per
                      certificate type and to drop restrictions where
                      needed (such as for S/MIME with shared mail
                      providers).</span></p>
                </div>
                <div>
                  <p class="x_MsoNormal"> </p>
                </div>
                <div>
                  <p class="x_MsoNormal"><span style="font-size:12.0pt;
                      color:black">I'm currently not in favor of having
                      a new S/MIME specific property, and one for client
                      certs, document signing, etc. as where does it
                      stop. It would also not allow to have separate
                      `iodef` settings for example (or only enable them
                      for TLS).</span></p>
                </div>
                <div>
                  <p class="x_MsoNormal"> </p>
                </div>
                <div>
                  <p class="x_MsoNormal"><span style="color:black">We
                      could define one new parameter to define the
                      certificate type, the standard already has a
                      reversed policy property which was used in the
                      draft RFC to limit issuance on policy OID. </span><span
                      style="font-size:12.0pt; color:black">Rob pointed
                      me at the draft CAA specification that had
                      a policy property value instead of a CA domain
                      name.</span></p>
                </div>
                <div>
                  <p class="x_MsoNormal"><span style="font-size:12.0pt;
                      color:black"><a
href="https://urldefense.com/v3/__https://datatracker.ietf.org/doc/html/draft-hallambaker-donotissue-04*section-3.1.2__;Iw!!FJ-Y8qCqXTj2!IBoj6xSTMxPkjo7rD0Gkn1l3AXVVMPwz8K-fe_d1vIOudW99epByL8XEpO9PYRLC7GtlLred0A$"
                        moz-do-not-send="true">https://datatracker.ietf.org/doc/html/draft-hallambaker-donotissue-04#section-3.1.2</a></span></p>
                </div>
                <div>
                  <p class="x_MsoNormal"><span style="font-size:12.0pt;
                      color:black"> </span></p>
                </div>
                <div>
                  <p class="x_MsoNormal"><span style="font-size:12.0pt;
                      color:black">This could work with cabforum defined
                      OID's, the advantage from using OID's is that it
                      would support an OID prefix, and it would be
                      easier for CAs to enforce. But it's not very user
                      friendly..?</span></p>
                </div>
                <div>
                  <p class="x_MsoNormal"><span style="font-size:12.0pt;
                      color:black"> </span></p>
                </div>
                <div>
                  <p class="x_MsoNormal"><span style="font-size:12.0pt;
                      color:black">This would allow:
                    </span></p>
                  <div>
                    <p class="x_MsoNormal"><span
                        style="font-size:10.5pt; font-family:"Segoe
                        UI",sans-serif; color:black"> </span></p>
                  </div>
                  <div>
                    <pre style="break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   $ORIGIN example.com</span><span style="color:black"></span></pre>
                    <pre style="break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issue "ca1.example.net; policy=2.23.140.1"   // Only cabforum</span><span style="color:black"></span></pre>
                    <pre style="break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issue "ca2.example.net; policy=2.23.140.1.5" // SMIME</span><span style="color:black"></span></pre>
                    <pre style="break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issue "ca3.example.net; policy=2.23.140.1.2" // DV, OV, IV</span><span style="color:black"></span></pre>
                    <pre style="break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issue "ca4.example.net; policy=2.23.140.1.1" // EV</span><span style="color:black"></span></pre>
                  </div>
                  <p class="x_MsoNormal"><span style="font-size:12.0pt;
                      color:black"> </span></p>
                </div>
                <div>
                  <p class="x_MsoNormal"><span style="font-size:12.0pt;
                      color:black"> </span></p>
                </div>
                <div>
                  <p class="x_MsoNormal"><span style="font-size:12.0pt;
                      color:black">For <span style="background:white">
                        user </span>friendliness, we could define a new
                      parameter such as `type` with the same intention
                      but based on a name value:</span></p>
                </div>
                <div>
                  <p class="x_MsoNormal"><span style="font-size:12.0pt;
                      color:black"> </span></p>
                </div>
                <div>
                  <p class="x_MsoNormal"><span style="font-size:12.0pt;
                      color:black">This would allow:</span></p>
                </div>
                <div>
                  <p class="x_MsoNormal"> </p>
                </div>
                <div>
                  <pre style="break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   $ORIGIN example.com</span></pre>
                  <pre><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issue "ca1.example.net"</span></pre>
                  <pre style="break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issue "ca2.example.net; type=smime"</span></pre>
                  <pre style="break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issue "ca3.example.net; type=tls"</span></pre>
                  <pre style="break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issue "ca4.example.net; type=tls-ev"</span></pre>
                  <pre><span style="font-family:"Calibri",sans-serif; color:#201F1E; background:white"> </span></pre>
                  <pre style="break-before:page"><span style="font-size:12.0pt; font-family:"Calibri",sans-serif; color:black">Where:</span></pre>
                  <pre style="margin-left:.5in; text-indent:-.25in; mso-list:l0 level1 lfo1; break-before:page"><span style="font-family:Symbol"><span style="mso-list:Ignore">·<span style="font:7.0pt "Times New Roman"">        </span></span></span><span style="font-size:12.0pt; font-family:"Calibri",sans-serif; color:black">ca1 could issue any type of certificate not explicitly specified (so no S/MIME, or TLS certificates in this example)</span></pre>
                  <pre style="margin-left:.5in; text-indent:-.25in; mso-list:l0 level1 lfo1"><span style="font-family:Symbol"><span style="mso-list:Ignore">·<span style="font:7.0pt "Times New Roman"">        </span></span></span><span style="font-size:12.0pt; font-family:"Calibri",sans-serif; color:black">ca2 could issue only smime certificates</span></pre>
                  <pre style="margin-left:.5in; text-indent:-.25in; mso-list:l0 level1 lfo1; break-before:page"><span style="font-family:Symbol"><span style="mso-list:Ignore">·<span style="font:7.0pt "Times New Roman"">        </span></span></span><span style="font-size:12.0pt; font-family:"Calibri",sans-serif; color:black">ca3 could issue any type of TLS certificate except for EV</span></pre>
                  <pre style="margin-left:.5in; text-indent:-.25in; mso-list:l0 level1 lfo1; break-before:page"><span style="font-family:Symbol"><span style="mso-list:Ignore">·<span style="font:7.0pt "Times New Roman"">        </span></span></span><span style="font-size:12.0pt; font-family:"Calibri",sans-serif; color:black">ca4 could issue only EV TLS certificates</span></pre>
                  <pre style="break-before:page"> </pre>
                  <pre style="break-before:page"><span style="font-size:12.0pt; font-family:"Calibri",sans-serif; color:black">Alternatively, we could define two parameters, one for the certificate type and one for the assurance level, this would give:</span></pre>
                  <pre style="break-before:page"> </pre>
                  <pre style="background:white; break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   $ORIGIN example.com</span></pre>
                  <pre style="background:white"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issue "ca1.example.net"</span><span style="font-size:10.5pt; color:black"></span></pre>
                  <pre style="background:white; break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issue "ca2.example.net; type=smime"</span><span style="color:black"></span></pre>
                  <pre style="background:white; break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issue "ca3.example.net; type=tls"</span><span style="color:black"></span></pre>
                  <pre style="background:white; break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issue "ca4.example.net; type=tls; level=ev;"</span><span style="color:black"></span></pre>
                  <p class="x_MsoNormal"> </p>
                </div>
                <div>
                  <p class="x_MsoNormal"><span style="font-size:12.0pt;
                      color:black">The challenge for all these methods
                      is how do we drop CAA limitations, as we want
                      something like:</span></p>
                </div>
                <div>
                  <p class="x_MsoNormal"><span style="font-size:12.0pt;
                      color:black"> </span></p>
                </div>
                <div>
                  <pre style="background:white; break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   $ORIGIN example.com</span></pre>
                  <pre style="background:white"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issue "ca1.example.net"</span><span style="font-size:10.5pt; font-family:"Calibri",sans-serif; color:black"></span></pre>
                  <pre style="break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black; background:white">   .       CAA 0 issue "*; type=smime"</span><span style="color:black; background:white"></span></pre>
                  <pre style="background:white; break-before:page"><span style="font-size:10.5pt; font-family:"Calibri",sans-serif; color:black"> </span></pre>
                  <pre style="break-before:page"><span style="font-size:12.0pt; font-family:"Calibri",sans-serif; color:black">But that is not allowed by the RFC.</span></pre>
                  <pre style="break-before:page"> </pre>
                  <pre style="break-before:page"><span style="font-size:12.0pt; font-family:"Calibri",sans-serif; color:black">If we would define a separate property per certificate type but follow the RFC and honoring the inheritance, we would still have the same challenge of stopping the inheritance .</span></pre>
                  <pre style="break-before:page"> </pre>
                  <pre style="background:white; break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   $ORIGIN example.com</span></pre>
                  <pre style="background:white"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issue "ca1.example.net"</span><span style="font-size:10.5pt; color:black"></span></pre>
                  <pre style="background:white; break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issuesmime "ca2.example.net"</span><span style="color:black"></span></pre>
                  <pre style="background:white; break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issuetls "ca3.example.net"</span><span style="color:black"></span></pre>
                  <pre style="background:white; break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issuetlsev "ca4.example.net"</span><span style="color:black"></span></pre>
                  <pre> </pre>
                  <pre style="break-before:page"><span style="font-size:12.0pt; font-family:"Calibri",sans-serif; color:black">Maybe we could simply create an 'unrestricted' parameter to overrule the RFC?:</span></pre>
                  <pre style="break-before:page"> </pre>
                  <pre style="background:white; break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   $ORIGIN example.com</span></pre>
                  <pre style="background:white"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issue "ca1.example.net"</span><span style="font-size:10.5pt; color:black"></span></pre>
                  <pre style="background:white; break-before:page"><span style="font-size:12.0pt; font-family:Consolas; color:black">   .       CAA 0 issue "; type=smime; unrestricted=true"</span><span style="color:black"></span></pre>
                  <pre> </pre>
                </div>
                <div>
                  <p class="x_MsoNormal"><span style="font-size:12.0pt;
                      color:black">Paul</span></p>
                </div>
                <div>
                  <p class="x_MsoNormal"><span style="font-size:12.0pt;
                      color:black"> </span></p>
                </div>
                <div class="x_MsoNormal" style="text-align:center"
                  align="center">
                  <hr width="98%" size="2" align="center">
                </div>
                <div id="x_divRplyFwdMsg">
                  <p class="x_MsoNormal"><b><span style="color:black">From:</span></b><span
                      style="color:black"> Smcwg-public <<a
                        href="mailto:smcwg-public-bounces@cabforum.org"
                        moz-do-not-send="true">smcwg-public-bounces@cabforum.org</a>>
                      on behalf of Tim Hollebeek via Smcwg-public <<a
                        href="mailto:smcwg-public@cabforum.org"
                        moz-do-not-send="true">smcwg-public@cabforum.org</a>><br>
                      <b>Sent:</b> Monday, October 26, 2020 15:25<br>
                      <b>To:</b> Neil Dunbar <<a
                        href="mailto:ndunbar@trustcorsystems.com"
                        moz-do-not-send="true">ndunbar@trustcorsystems.com</a>>;
                      SMIME Certificate Working Group <<a
                        href="mailto:smcwg-public@cabforum.org"
                        moz-do-not-send="true">smcwg-public@cabforum.org</a>><br>
                      <b>Subject:</b> [EXTERNAL]Re: [Smcwg-public] CAA
                      and S/MIME</span> </p>
                  <div>
                    <p class="x_MsoNormal"> </p>
                  </div>
                </div>
                <div>
                  <div>
                    <p class="x_xmsonormal">This is how I feel about the
                      issue.  CAA is potentially an interesting
                      improvement to the S/MIME ecosystem, but the
                      current tags and implementation were meant for
                      TLS, and shouldn’t be reused.</p>
                    <p class="x_xmsonormal"> </p>
                    <p class="x_xmsonormal">The RFC has an extension
                      mechanism which can easily be used to add new tags
                      for S/MIME issuance, and issuance of other kinds
                      of non-TLS certificates.</p>
                    <p class="x_xmsonormal"> </p>
                    <p class="x_xmsonormal">-Tim</p>
                    <p class="x_xmsonormal"> </p>
                    <div style="border:none; border-left:solid blue
                      1.5pt; padding:0in 0in 0in 4.0pt">
                      <div>
                        <div style="border:none; border-top:solid
                          #E1E1E1 1.0pt; padding:3.0pt 0in 0in 0in">
                          <p class="x_xmsonormal"><b>From:</b>
                            Smcwg-public <<a
                              href="mailto:smcwg-public-bounces@cabforum.org"
                              moz-do-not-send="true">smcwg-public-bounces@cabforum.org</a>>
                            <b>On Behalf Of </b>Neil Dunbar via
                            Smcwg-public<br>
                            <b>Sent:</b> Monday, October 26, 2020 6:03
                            AM<br>
                            <b>To:</b> <a
                              href="mailto:smcwg-public@cabforum.org"
                              moz-do-not-send="true">smcwg-public@cabforum.org</a><br>
                            <b>Subject:</b> Re: [Smcwg-public] CAA and
                            S/MIME</p>
                        </div>
                      </div>
                      <p class="x_xmsonormal"> </p>
                      <p> </p>
                      <div>
                        <p class="x_xmsonormal">On 24/10/2020 16:21,
                          Stephen Davidson via Smcwg-public wrote:</p>
                      </div>
                      <blockquote style="margin-top:5.0pt;
                        margin-bottom:5.0pt">
                        <p class="x_xmsonormal">Hello:</p>
                        <p class="x_xmsonormal"> </p>
                        <p class="x_xmsonormal">The topic of
                          Certification Authority Authorisation (CAA)
                          has arisen a number of times in relation to
                          the evolving S/MIME Baseline.</p>
                        <p class="x_xmsonormal">I highlight a discussion
                          on that subject related to the Mozilla policy:
                          <a
href="https://urldefense.com/v3/__https://github.com/mozilla/pkipolicy/issues/135__;!!FJ-Y8qCqXTj2!IBoj6xSTMxPkjo7rD0Gkn1l3AXVVMPwz8K-fe_d1vIOudW99epByL8XEpO9PYRLC7GspokrqcA$"
                            moz-do-not-send="true">
https://github.com/mozilla/pkipolicy/issues/135</a></p>
                        <p class="x_xmsonormal">A significant number of
                          email providers – such as gmail.com,
                          outlook.com, protonmail.com, and others – have
                          CAA records.</p>
                        <p class="x_xmsonormal"><br>
                          Questions for us to address later in our
                          discussions:</p>
                        <p class="x_xmsonormal"> </p>
                        <ul style="margin-top:0in" type="disc">
                          <li class="x_xmsolistparagraph"
                            style="margin-top:0in; margin-bottom:0in;
                            mso-list:l1 level1 lfo2">
                            Is CAA a desired requirement of the S/MIME
                            Baseline?</li>
                          <li class="x_xmsolistparagraph"
                            style="margin-top:0in; margin-bottom:0in;
                            mso-list:l1 level1 lfo2">
                            Should the S/MIME Baseline rely upon the
                            existing requirements stated in the TLS BR,
                            or is the S/MIME use case sufficiently
                            different to merit a separate CAA tag?</li>
                        </ul>
                        <p class="x_xmsonormal"
                          style="margin-bottom:12.0pt"> </p>
                      </blockquote>
                      <p>Generally, I'm a fan of allowing organisations
                        (however defined) to specify their policy
                        requirements for publicly trusted certificates
                        via CAA records; so I would say "yes", it is a
                        desired requirement of the S/MIME baseline. I
                        would certainly expect it to make its way into
                        the root program requirements at some point, and
                        having a pan-root program consensus on those
                        requirements beats having overlapping or
                        potentially conflicting requirements.</p>
                      <p>That said, I'm not a fan of ninja semantics
                        (changing the meaning of a deployed resource
                        where the deployer might not have considered its
                        eventual full scope) - it seems to me that the
                        "issue" and "issuewild" tags were framed with
                        TLS certificates in mind[*], and I think
                        extending "issue" to cover S/MIME could have
                        effects on domain owners which were not
                        expected. In other words, we would be saying to
                        them that all certificates are hereby covered,
                        without them having any means of expressing the
                        policy "I want CA X to issue TLS certificates,
                        but any CA could issue S/MIME certificates"; so
                        I'm less of a fan of reducing the expressive
                        potential of domain owners.</p>
                      <p>To that extent, I think that I'd prefer tags
                        like "issue-tls", "issue-tls-wildcard",
                        "issue-email", and so on, with similar semantics
                        which work over "issue" and "issuewild" right
                        now. Once those are in effect, then you could
                        extend the "issue" tag to mean "all
                        certificates" as a shorthand, while leaving
                        finer detailed policy expressions. However, that
                        goes further than anything the S/MIME WG could
                        reasonably pronounce upon. But "issue-email" to
                        cover S/MIME certs falls within its charter and
                        seems to have a clearer scope. There's even an
                        opportunity to allow domain owners to specify
                        the validation methods permissible for issuance,
                        but that's a whole different discussion.</p>
                      <p>Just my opinion, of course.</p>
                      <p>Neil</p>
                      <p>[*] Genuine question: would "issuewild" have
                        any meaning outside of TLS certificates?</p>
                    </div>
                  </div>
                </div>
              </div>
            </div>
            <br>
          </div>
          <br>
          <fieldset class="x_mimeAttachmentHeader"></fieldset>
          <pre class="x_moz-quote-pre">_______________________________________________
Smcwg-public mailing list
<a class="x_moz-txt-link-abbreviated" href="mailto:Smcwg-public@cabforum.org" moz-do-not-send="true">Smcwg-public@cabforum.org</a>
<a class="x_moz-txt-link-freetext" href="https://urldefense.com/v3/__https://lists.cabforum.org/mailman/listinfo/smcwg-public__;!!FJ-Y8qCqXTj2!IBoj6xSTMxPkjo7rD0Gkn1l3AXVVMPwz8K-fe_d1vIOudW99epByL8XEpO9PYRLC7GuY2WazZQ$" moz-do-not-send="true">https://lists.cabforum.org/mailman/listinfo/smcwg-public</a>
</pre>
        </blockquote>
        <br>
      </div>
    </blockquote>
    <br>
  </body>
</html>