<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE rfc [
<!ENTITY nbsp "&#160;">
]>
<rfc xmlns:xi="http://www.w3.org/2001/XInclude" category="info" docName="draft-condrey-rats-pop-examples-00" ipr="trust200902" submissionType="independent" xml:lang="en" version="3">

  <front>
    <title abbrev="PoP Examples">Proof of Process: Worked Examples</title>

    <seriesInfo name="Internet-Draft" value="draft-condrey-rats-pop-examples-00"/>

    <author fullname="David Condrey" initials="D." surname="Condrey">
      <organization>Writerslogic Inc</organization>
      <address>
        <postal>
          <country>United States</country>
        </postal>
        <email>david@writerslogic.com</email>
      </address>
    </author>

    <date year="2026" month="February" day="6"/>

    <area>Security</area>
    <workgroup>Remote ATtestation procedureS</workgroup>

    <keyword>CBOR</keyword>
    <keyword>examples</keyword>
    <keyword>attestation</keyword>
    <keyword>evidence</keyword>

    <abstract>
      <t>
        This document provides worked examples demonstrating the Proof of
        Process Evidence format and Attestation Results. Examples include
        minimal Evidence packets, multi-checkpoint scenarios, jitter seal
        verification, VDF causality chains, and salt mode configurations.
        This companion document supplements the main Proof of Process
        specification with practical reference implementations.
      </t>
    </abstract>
  </front>

  <middle>
    <section anchor="introduction">
      <name>Introduction</name>

      <t>
        This document provides worked examples of Proof of Process (PoP)
        Evidence packets and Attestation Results as defined in
        <xref target="I-D.condrey-rats-pop"/>. All examples use CBOR
        diagnostic notation <xref target="RFC8949"/> Section G, with
        comments (/ ... /) to annotate fields. Integer keys are used as
        defined in the companion CDDL schema.
      </t>

      <t>
        These examples are informative. The normative schema is defined in
        the CDDL schema document. Implementers SHOULD validate their
        implementations against test vectors derived from these examples.
      </t>

      <section anchor="conventions">
        <name>Requirements Language</name>
        <t>
          The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
          "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and
          "OPTIONAL" in this document are to be interpreted as described in
          BCP 14 <xref target="RFC2119"/> <xref target="RFC8174"/> when, and
          only when, they appear in all capitals, as shown here.
        </t>
      </section>
    </section>

    <section anchor="example-minimal">
      <name>Minimal Evidence Packet (Basic Tier)</name>

      <t>
        This example demonstrates a minimal Basic tier Evidence packet
        with a single checkpoint. The document is approximately 500
        characters, representing a short paragraph written in a single
        authoring session.
      </t>

      <t>
        Scenario: An author writes a brief memo over approximately 3
        minutes. The Attesting Environment captures one checkpoint at
        the end of the session.
      </t>

      <sourcecode type="cbor-diag"><![CDATA[
1347571280({
  1: 1,
  2: "https://writerslogic.com/rats/eat/profile/pop/1.0",
  3: h'550e8400e29b41d4a716446655440000',
  4: 1(1706745600),

  5: {
    1: {
      1: 1,
      2: h'e3b0c44298fc1c149afbf4c8996fb924
          27ae41e4649b934ca495991b7852b855'
    },
    2: "memo.txt",
    3: 487,
    4: 478,
    5: 0
  },

  6: [
    {
      1: 0,
      2: h'6ba7b8109dad11d180b400c04fd430c8',
      3: 1(1706745780),

      4: {
        1: 1,
        2: h'e3b0c44298fc1c149afbf4c8996fb924
            27ae41e4649b934ca495991b7852b855'
      },

      5: 478,
      6: 87,

      7: {
        1: 478, 2: 12, 3: 47, 4: 8,
        5: 3, 6: 0, 7: 0, 8: 14, 9: 180.0
      },

      8: {
        1: 1,
        2: h'00000000000000000000000000000000
            00000000000000000000000000000000'
      },

      9: {
        1: 1,
        2: h'a7ffc6f8bf1ed76651c14756a061d662
            f580ff4de43b49fa82d80a4b80f8434a'
      },

      10: {
        1: 1,
        2: {1: 1, 2: 8500000},
        3: h'9f86d081884c7d659a2feaa0c55ad015
            a3bf4f1b2b0b822cd15d6c15b0f00a08',
        4: h'2c624232cdd221771294dfbb310aca00
            0a0df6ac8b66b696d90ef06fdefb64a3',
        5: h'',
        6: 180.0,
        7: 1530000000,
        8: {
          1: 8500000,
          2: 1(1706745600),
          3: h'deadbeef...',
          4: h'cafebabe...',
          5: "MacBook Pro M3"
        }
      },

      11: {
        1: {
          1: 1,
          2: h'7d865e959b2466918c9863afca942d0f
              b89d7c9ac0c99bafc3749504ded97730'
        },
        2: (1, 2, 3),
        3: {
          1: 423,
          2: [
            {1: 0, 2: 50, 3: 12},
            {1: 50, 2: 100, 3: 89},
            {1: 100, 2: 200, 3: 156},
            {1: 200, 2: 500, 3: 98},
            {1: 500, 2: 1000, 3: 34},
            {1: 1000, 2: 2000, 3: 18},
            {1: 2000, 2: 5000, 3: 12},
            {1: 5000, 2: 4294967295, 3: 4}
          ],
          3: 2.78,
          4: []
        },
        4: h'b94d27b9934d3e08a52e52d7da7dabfa
            c484efe37a5380ee9088f7ace2efcde9'
      },

      12: h'73475cb40a568e8da8a045ced110137e
          159f890ac4da883b6b17dc651b3a8049'
    }
  ]
})
]]></sourcecode>

      <t>
        Key observations:
      </t>

      <ul>
        <li>
          The Basic tier contains only the required checkpoint chain.
          No optional sections (presence, forensics, etc.) are included.
        </li>
        <li>
          The genesis checkpoint has prev-hash of 32 zero bytes.
        </li>
        <li>
          The VDF proves at least 180 seconds elapsed (1.53B iterations
          at calibrated 8.5M iterations/second).
        </li>
        <li>
          Jitter binding shows 423 timing samples with realistic
          distribution peaking in the 100-200ms range.
        </li>
        <li>
          Estimated entropy is 2.78 bits, below the recommended 32-bit
          threshold for Standard tier but acceptable for Basic tier.
        </li>
      </ul>
    </section>

    <section anchor="example-multi-checkpoint">
      <name>Multi-Checkpoint Evidence (Standard Tier)</name>

      <t>
        This example demonstrates a Standard tier Evidence packet with
        three checkpoints showing document evolution. The document grows
        from 100 to 500 to 1200 characters across the checkpoints,
        representing a typical drafting process with revisions.
      </t>

      <t>
        Scenario: An author writes a short essay over 45 minutes with
        two natural breaks where checkpoints are captured. The evidence
        includes presence challenges.
      </t>

      <sourcecode type="cbor-diag"><![CDATA[
1347571280({
  1: 1,
  2: "https://writerslogic.com/rats/eat/profile/pop/1.0",
  3: h'123e4567e89b12d3a456426614174000',
  4: 1(1706832000),

  5: {
    1: {1: 1, 2: h'abcd1234...'},
    3: 1247,
    4: 1203,
    5: 0
  },

  6: [
    {
      1: 0,
      2: h'a1b2c3d4e5f6a7b8c9d0e1f2a3b4c5d6',
      3: 1(1706832600),
      4: {1: 1, 2: h'1111aaaa...'},
      5: 103,
      6: 19,
      7: {
        1: 103, 2: 8, 3: 22, 4: 5,
        5: 2, 6: 0, 7: 0, 8: 6, 9: 600.0
      },
      8: {1: 1, 2: h'00000000...'},
      9: {1: 1, 2: h'2222bbbb...'},
      10: {
        1: 1,
        2: {1: 1, 2: 8500000},
        3: h'input0...',
        4: h'output0...',
        5: h'',
        6: 600.0,
        7: 5100000000,
        8: {
          1: 8500000, 2: 1(1706832000),
          3: h'sig...', 4: h'nonce...'
        }
      },
      11: {
        1: {1: 1, 2: h'jitter0...'},
        2: (1, 2),
        3: {1: 156, 2: (...), 3: 2.45},
        4: h'mac0...'
      },
      12: h'chainmac0...'
    },

    {
      1: 1,
      2: h'b2c3d4e5f6a7b8c9d0e1f2a3b4c5d6e7',
      3: 1(1706833800),
      4: {1: 1, 2: h'3333cccc...'},
      5: 512,
      6: 94,
      7: {
        1: 423, 2: 14, 3: 87, 4: 12,
        5: 6, 6: 0, 7: 0, 8: 23, 9: 1200.0
      },
      8: {1: 1, 2: h'2222bbbb...'},
      9: {1: 1, 2: h'4444dddd...'},
      10: {
        1: 1,
        2: {1: 1, 2: 8500000},
        3: h'H(output0 || content-hash{1} || jitter-commitment{1})',
        4: h'output1...',
        5: h'',
        6: 1200.0,
        7: 10200000000
      },
      11: {
        1: {1: 1, 2: h'jitter1...'},
        2: (1, 2, 3),
        3: {1: 298, 2: (...), 3: 2.67},
        4: h'mac1...'
      },
      12: h'chainmac1...'
    },

    {
      1: 2,
      2: h'c3d4e5f6a7b8c9d0e1f2a3b4c5d6e7f8',
      3: 1(1706835000),
      4: {1: 1, 2: h'abcd1234...'},
      5: 1203,
      6: 218,
      7: {
        1: 712, 2: 21, 3: 134, 4: 18,
        5: 8, 6: 0, 7: 0, 8: 45, 9: 1200.0
      },
      8: {1: 1, 2: h'4444dddd...'},
      9: {1: 1, 2: h'5555eeee...'},
      10: {
        1: 1,
        2: {1: 1, 2: 8500000},
        3: h'H(output1 || content-hash{2} || jitter-commitment{2})',
        4: h'output2...',
        5: h'',
        6: 1200.0,
        7: 10200000000
      },
      11: {
        1: {1: 1, 2: h'jitter2...'},
        2: (1, 2, 3, 4),
        3: {1: 387, 2: (...), 3: 2.89},
        4: h'mac2...'
      },
      12: h'chainmac2...'
    }
  ],

  10: {
    1: [
      {
        1: 1(1706832900),
        2: 1,
        3: 847,
        4: true,
        5: h'challenge-nonce-1'
      },
      {
        1: 1(1706834400),
        2: 2,
        3: 2134,
        4: true,
        5: h'challenge-nonce-2'
      }
    ],
    2: {
      1: 2,
      2: 2,
      3: 2,
      4: 1490
    }
  }
})
]]></sourcecode>

      <t>
        VDF entanglement across checkpoints:
      </t>

      <artwork><![CDATA[
VDF Chain Causality:
+---------------+     +---------------+     +---------------+
| Checkpoint 0  |     | Checkpoint 1  |     | Checkpoint 2  |
|---------------|     |---------------|     |---------------|
| VDF_input{0}  |     | VDF_input{1}  |     | VDF_input{2}  |
|   = H(session |     |   = H(        |     |   = H(        |
|       entropy |     |     output{0} |     |     output{1} |
|    || content |     |  || content{1}|     |  || content{2}|
|    || jitter) |     |  || jitter{1})|     |  || jitter{2})|
|---------------|     |---------------|     |---------------|
| VDF_output{0} |---->| VDF_output{1} |---->| VDF_output{2} |
+---------------+     +---------------+     +---------------+

To backdate checkpoint 1, adversary must:
(1) Compute content that hashes to content-hash{1}
(2) Generate jitter that commits to jitter-commitment{1}
(3) Recompute VDF_output{1} from new VDF_input{1}
(4) Recompute VDF_output{2} (depends on output{1})
(5) Complete steps 3-4 before external anchor confirms state
]]></artwork>
    </section>

    <section anchor="example-jitter-verification">
      <name>Jitter Seal Verification Example</name>

      <t>
        This example shows the complete jitter seal verification process,
        including histogram data, entropy calculation, and binding MAC
        computation.
      </t>

      <section anchor="jitter-sample-data">
        <name>Sample Histogram Data</name>

        <sourcecode type="cbor-diag"><![CDATA[
jitter-binding = {
  1: {
    1: 1,
    2: h'b94d27b9934d3e08a52e52d7da7dabfa
        c484efe37a5380ee9088f7ace2efcde9'
  },

  2: (1, 2, 3),

  3: {
    1: 842,
    2: [
      {1: 0,    2: 50,   3: 24},
      {1: 50,   2: 100,  3: 178},
      {1: 100,  2: 200,  3: 312},
      {1: 200,  2: 500,  3: 196},
      {1: 500,  2: 1000, 3: 68},
      {1: 1000, 2: 2000, 3: 36},
      {1: 2000, 2: 5000, 3: 22},
      {1: 5000, 2: 4294967295, 3: 6}
    ],
    3: 2.54,
    4: []
  },

  4: h'73475cb40a568e8da8a045ced110137e
      159f890ac4da883b6b17dc651b3a8049',

  5: (87, 134, 112, 98, 203, 156, 89, 167, 1243, 78, ...)
}
]]></sourcecode>
      </section>

      <section anchor="entropy-calculation">
        <name>Entropy Calculation Walkthrough</name>

        <t>
          Shannon entropy is calculated from the histogram distribution:
        </t>

        <artwork><![CDATA[
Given histogram counts: (24, 178, 312, 196, 68, 36, 22, 6)
Total samples: 842

(1) Calculate probabilities p(i) = count(i) / total
  p{0} = 24/842  = 0.0285
  p{1} = 178/842 = 0.2114
  p{2} = 312/842 = 0.3705
  p{3} = 196/842 = 0.2328
  p{4} = 68/842  = 0.0808
  p{5} = 36/842  = 0.0428
  p{6} = 22/842  = 0.0261
  p{7} = 6/842   = 0.0071

(2) Calculate Shannon entropy H = -sum(p{i} * log2(p{i}))
  H = -(0.0285 * log2(0.0285)   = 0.148
      + 0.2114 * log2(0.2114)   = 0.467
      + 0.3705 * log2(0.3705)   = 0.531
      + 0.2328 * log2(0.2328)   = 0.481
      + 0.0808 * log2(0.0808)   = 0.295
      + 0.0428 * log2(0.0428)   = 0.197
      + 0.0261 * log2(0.0261)   = 0.137
      + 0.0071 * log2(0.0071))  = 0.050

  H = 2.306 bits (per sample from 8-bucket distribution)

(3) Estimated total entropy
  Total entropy bits = H * log2(sample_count)
  = 2.306 * log2(842)
  = 2.306 * 9.72
  = 22.4 bits (approximate)

  Reported: 2.54 bits (average per-sample entropy)
]]></artwork>
      </section>

      <section anchor="binding-mac-computation">
        <name>Binding MAC Computation</name>

        <artwork><![CDATA[
binding-mac = HMAC-SHA256(
    key = checkpoint-chain-key,
    message = entropy-commitment ||
              CBOR(sources) ||
              CBOR(summary) ||
              prev-checkpoint-hash
)

Where:
  checkpoint-chain-key = session-derived 256-bit key
  entropy-commitment   = h'b94d27b9...' (32 bytes)
  CBOR(sources)        = 83 01 02 03 (4 bytes: array of 3 uints)
  CBOR(summary)        = A4 01 ... (variable length map)
  prev-checkpoint-hash = h'...' (32 bytes)
]]></artwork>
      </section>

      <section anchor="verifier-jitter-checks">
        <name>Verifier Checks</name>

        <t>
          A Verifier performs the following checks on the jitter seal:
        </t>

        <sourcecode type="pseudocode"><![CDATA[
def verify_jitter_seal(jitter_binding, prev_hash, chain_key):
    # (1) Structural validation
    assert all_required_fields_present(jitter_binding)
    assert len(jitter_binding.sources) >= 1

    # (2) Recompute entropy-commitment (if raw-intervals disclosed)
    if jitter_binding.raw_intervals is not None:
        expected_commitment = sha256(
            concat_as_uint32_le(jitter_binding.raw_intervals)
        )
        commitment = jitter_binding.entropy_commitment
        assert commitment == expected_commitment

    # (3) Recompute histogram (if raw-intervals disclosed)
    if jitter_binding.raw_intervals is not None:
        computed_histogram = bucket_intervals(
            jitter_binding.raw_intervals,
            boundaries=(0, 50, 100, 200, 500, 1000, 2000, 5000)
        )
        assert histograms_consistent(
            computed_histogram,
            jitter_binding.summary.timing_histogram
        )

    # (4) Verify binding MAC
    expected_mac = hmac_sha256(
        key=chain_key,
        message=jitter_binding.entropy_commitment.value +
                cbor_encode(jitter_binding.sources) +
                cbor_encode(jitter_binding.summary) +
                prev_hash
    )
    assert jitter_binding.binding_mac == expected_mac

    # (5) Entropy threshold check
    assert jitter_binding.summary.entropy_bits >= MIN_THRESHOLD

    # (6) Sample count plausibility
    assert jitter_binding.summary.sample_count >= 10

    return VERIFIED
]]></sourcecode>
      </section>
    </section>

    <section anchor="example-vdf-causality">
      <name>VDF Causality Example</name>

      <t>
        This example demonstrates VDF input computation and why backdating
        requires recomputation of the entire subsequent chain.
      </t>

      <section anchor="vdf-input-computation">
        <name>VDF Input Computation for Checkpoint N</name>

        <sourcecode type="cbor-diag"><![CDATA[
VDF_input{N} = SHA256(
    VDF_output{N-1} ||
    content-hash{N} ||
    jitter-commitment{N} ||
    uint32_le(sequence{N})
)

VDF_input{2} = SHA256(
    h'output1...'
 || h'abcd1234...'
 || h'jitter2...'
 || h'02000000'
)

VDF_output{2} = SHA256^10200000000(VDF_input{2})
]]></sourcecode>
      </section>

      <section anchor="backdating-cost">
        <name>Why Backdating Requires Recomputation</name>

        <artwork><![CDATA[
Adversary Goal: Insert a fake checkpoint between checkpoints 0 and 1

Attempt: Create fake checkpoint 0.5 with:
  - content-hash{0.5} = hash of backdated content
  - jitter-commitment{0.5} = fabricated timing data
  - sequence{0.5} = 1 (bumping original checkpoint 1 to sequence 2)

Problem: This changes VDF_input{1}:
  VDF_input{1}_original = H(VDF_output{0} || content{1} || jitter{1})
  VDF_input{1}_fake     = H(VDF_output{0.5} || content{1} || jitter{1})

Since VDF_output{0.5} != VDF_output{0}, the adversary must:

(1) Compute VDF_output{0.5} from VDF_input{0.5}
   Cost: ~1200 seconds (cannot parallelize)

(2) Recompute VDF_output{1} from new VDF_input{1}
   Cost: ~1200 seconds (cannot parallelize)

(3) Recompute VDF_output{2} from new VDF_input{2}
   Cost: ~1200 seconds (cannot parallelize)

Total minimum time: 3600 seconds = 1 hour
(Cannot be reduced by parallel computation)

If external anchor confirmed checkpoint 2 at T_anchor,
adversary must complete all recomputation before T_anchor.
Any attempt to backdate beyond anchor is impossible.
]]></artwork>
      </section>

      <section anchor="calibration-cross-check">
        <name>Calibration Cross-Check</name>

        <sourcecode type="pseudocode"><![CDATA[
def verify_vdf_duration_claim(checkpoint, calibration):
    # Extract values
    iterations = checkpoint.vdf_proof.iterations
    claimed_duration = checkpoint.vdf_proof.claimed_duration
    calibration_rate = calibration.calibration_iterations

    # Compute minimum possible duration
    min_duration = iterations / calibration_rate

    # Allow 10% tolerance for measurement variance
    tolerance = 1.1

    # Claimed duration must be at least min_duration
    if claimed_duration < (min_duration / tolerance):
        return INVALID("Claimed duration impossibly short")

    # Claimed duration should not be excessively long
    if claimed_duration > (min_duration * 10):
        return WARNING("Claimed duration suspiciously long")

    return VALID

# Example:
# iterations = 10,200,000,000
# calibration_rate = 8,500,000 / second
# min_duration = 10.2B / 8.5M = 1200 seconds
# claimed_duration = 1200 seconds: VALID
]]></sourcecode>
      </section>
    </section>

    <section anchor="example-absence-claim">
      <name>Absence Claim Example</name>

      <t>
        This example demonstrates both chain-verifiable and monitoring-
        dependent absence claims, showing how verifiers prove claims
        from checkpoint data.
      </t>

      <section anchor="chain-verifiable-example">
        <name>Chain-Verifiable Claim: max-single-delta-chars</name>

        <t>
          Claim: No single checkpoint added more than 500 characters.
        </t>

        <sourcecode type="cbor-diag"><![CDATA[
absence-claim = {
  1: 1,
  2: {1: 500},
  3: {
    1: 1,
    2: {
      1: (0, 2),
      2: 423
    }
  },
  4: {
    1: 1,
    2: []
  }
}
]]></sourcecode>

        <t>
          Verifier proof procedure:
        </t>

        <sourcecode type="pseudocode"><![CDATA[
def verify_max_single_delta_chars(evidence, threshold):
    # (1) Verify chain integrity
    assert verify_chain_hashes(evidence.checkpoints)
    assert verify_vdf_linkage(evidence.checkpoints)

    # (2) Extract max delta from checkpoint data
    max_chars_added = 0
    for checkpoint in evidence.checkpoints:
        delta = checkpoint.delta
        max_chars_added = max(max_chars_added, delta.chars_added)

    # (3) Compare against threshold
    if max_chars_added <= threshold:
        return PROVEN(
            observed=max_chars_added,
            threshold=threshold,
            confidence="proven"
        )
    else:
        return FAILED(
            observed=max_chars_added,
            threshold=threshold
        )
]]></sourcecode>
      </section>

      <section anchor="monitoring-dependent-example">
        <name>Monitoring-Dependent Claim: max-paste-event-chars</name>

        <t>
          Claim: No paste event inserted more than 200 characters.
        </t>

        <sourcecode type="cbor-diag"><![CDATA[
absence-claim = {
  1: 16,
  2: {1: 200},
  3: {
    1: 2,
    2: {
      1: (0, 2),
      2: 0,
      3: 0.0
    }
  },
  4: {
    1: 2,
    2: (
      "Requires trust in clipboard monitoring",
      "Coverage fraction: 0.98"
    )
  },
  5: {
    1: 2,
    2: "macOS NSPasteboard notifications monitored continuously",
    3: true
  }
}
]]></sourcecode>

        <t>
          Trust chain for monitoring-dependent claim:
        </t>

        <artwork><![CDATA[
Trust Chain Analysis:

(1) AE Trust Target: os-reported-events (2)
   - Requires: OS correctly reports clipboard access
   - macOS: NSPasteboard change notifications
   - Trustworthiness: Moderate (depends on OS integrity)

(2) Verification Status: true
   - Cross-reference: hardware-section contains SE attestation
   - SE attests: witnessd binary hash, measurement time
   - This increases confidence that AE was unmodified

(3) Monitoring Coverage: 98%
   - monitoring-coverage.coverage-fraction = 0.98
   - 2% gap when app was backgrounded
   - Caveat: paste during gap would be undetected

(4) Resulting Confidence: HIGH (level 2)
   - Not PROVEN (would require trustless verification)
   - HIGH because: HW attestation + high coverage + OS events

Relying Party Decision:
  - For academic submission: HIGH confidence acceptable
  - For legal proceeding: May require additional corroboration
]]></artwork>
      </section>
    </section>

    <section anchor="example-attestation-result">
      <name>Attestation Result Example (.war)</name>

      <t>
        This example shows a Verifier's Attestation Result after
        appraising the multi-checkpoint Evidence packet from
        <xref target="example-multi-checkpoint"/>.
      </t>

      <sourcecode type="cbor-diag"><![CDATA[
1463894560({
  1: 1,
  2: h'123e4567e89b12d3a456426614174000',
  3: 1(1706840000),

  4: 2,

  5: 0.78,

  6: [
    {
      1: 6,
      2: true,
      3: "Sequence numbers 0,1,2 consecutive",
      4: 1
    },
    {
      1: 7,
      2: true,
      3: "All prev-hash values match prior checkpoint-hash",
      4: 1
    },
    {
      1: 4,
      2: true,
      3: "Total VDF time: 3000 seconds (threshold: 2700)",
      4: 1
    },
    {
      1: 8,
      2: true,
      3: "Cumulative entropy: 7.92 bits (threshold: 6.0)",
      4: 1
    },
    {
      1: 16,
      2: true,
      3: "No paste events detected (threshold: 500)",
      4: 2
    },
    {
      1: 100,
      2: true,
      3: "2/2 challenges passed, median response 1490ms",
      4: 2
    }
  ],

  7: 18(h'D28441A0A201260442313154...'),

  8: "WritersLogic Verification Service v2.1",

  9: {
    1: "2.1.0",
    2: "https://verify.writerslogic.com",
    4: "pop-standard-v1"
  },

  10: (
    "No hardware attestation in Evidence packet",
    "Monitoring coverage: 98%",
    "VDF calibration self-reported"
  )
})
]]></sourcecode>

      <t>
        Key aspects of the Attestation Result:
      </t>

      <ul>
        <li>
          <t>Verdict: likely-human (2)</t>
          <t>
            Based on: checkpoint chain integrity, realistic jitter
            distribution, presence challenges passed, no anomalies.
          </t>
        </li>

        <li>
          <t>Confidence: 0.78 (high range)</t>
          <t>
            Factors increasing: chain integrity proven, presence verified.
            Factors limiting: no hardware attestation, self-reported
            calibration.
          </t>
        </li>

        <li>
          <t>Caveats document limitations:</t>
          <t>
            Relying Parties can assess whether the caveats are acceptable
            for their use case. Academic submission may accept; legal
            proceeding may require additional evidence.
          </t>
        </li>

        <li>
          <t>Verifier signature enables trust chain:</t>
          <t>
            Relying Party trusts WritersLogic Verification Service.
            Signature proves Attestation Result came from that Verifier.
          </t>
        </li>
      </ul>
    </section>

    <section anchor="example-salt-modes">
      <name>Salt Mode Examples</name>

      <t>
        This example demonstrates the three salt modes and their
        verification flow differences. The same document is shown
        with each salt mode.
      </t>

      <section anchor="salt-unsalted">
        <name>Unsalted Mode (Default)</name>

        <sourcecode type="cbor-diag"><![CDATA[
document-ref = {
  1: {
    1: 1,
    2: h'e3b0c44298fc1c149afbf4c8996fb924
        27ae41e4649b934ca495991b7852b855'
  },
  2: "essay.txt",
  3: 4523,
  4: 4401,
  5: 0
}
]]></sourcecode>

        <t>
          Verification flow (unsalted):
        </t>

        <artwork><![CDATA[
Verifier has: document content, Evidence packet

(1) Compute document hash
  computed_hash = SHA256(document_content)
  = h'e3b0c442...'

(2) Compare with Evidence
  assert computed_hash == document_ref.content_hash.value
  assert computed_hash == checkpoints{-1}.content_hash.value

Result: VERIFIED
  - Anyone with the document can verify binding
  - No additional information needed
  - Document linkage is globally verifiable
]]></artwork>
      </section>

      <section anchor="salt-author">
        <name>Author-Salted Mode</name>

        <sourcecode type="cbor-diag"><![CDATA[
document-ref = {
  1: {
    1: 1,
    2: h'7f83b1657ff1fc53b92dc18148a1d65d
        fc2d4b1fa3d677284addd200126d9069'
  },
  2: "confidential-report.txt",
  3: 8934,
  4: 8712,
  5: 1,
  6: h'9f86d081884c7d659a2feaa0c55ad015
      a3bf4f1b2b0b822cd15d6c15b0f00a08'
}
]]></sourcecode>

        <t>
          Verification flow (author-salted):
        </t>

        <artwork><![CDATA[
Verifier has: document content, Evidence packet
Verifier needs: salt (provided out-of-band by author)

(1) Author provides salt to chosen verifier
  received_salt = h'deadbeefcafebabe1234567890abcdef'

(2) Verify salt matches commitment
  computed_commitment = SHA256(received_salt)
  assert computed_commitment == document_ref.salt_commitment

(3) Compute salted document hash
  computed_hash = SHA256(received_salt || document_content)

(4) Compare with Evidence
  assert computed_hash == document_ref.content_hash.value

Result: VERIFIED
  - Only verifiers who received salt can verify
  - Author controls who can verify document binding
  - Useful for: unpublished manuscripts, confidential documents
]]></artwork>
      </section>

      <section anchor="salt-escrowed">
        <name>Third-Party Escrowed Mode</name>

        <sourcecode type="cbor-diag"><![CDATA[
document-ref = {
  1: {
    1: 1,
    2: h'3a7bd3e2360a3d29eea436fcfb7e44c7
        35d117c42d1c1835420b6b9942dd4f1b'
  },
  3: 12045,
  4: 11823,
  5: 2,
  6: h'a591a6d40bf420404a011733cfb7b190
      d62c65bf0bcda32b57b277d9ad9f146e'
}
]]></sourcecode>

        <t>
          Verification flow (escrowed):
        </t>

        <artwork><![CDATA[
Verifier has: document content, Evidence packet
Verifier needs: salt (from escrow service)

(1) Request salt from escrow
  - Verifier contacts escrow service
  - Escrow verifies release conditions are met:
    * Legal subpoena
    * Author consent
    * Time-based release
    * Dispute resolution trigger
  - Escrow provides salt if conditions satisfied

(2-4) Same as author-salted verification

Use cases:
  - Litigation discovery: salt released upon court order
  - Embargo periods: salt released after publication date
  - Dispute resolution: salt released if authorship contested
  - Dead man's switch: salt released if author inactive
]]></artwork>
      </section>

      <section anchor="salt-comparison">
        <name>Salt Mode Comparison</name>

        <table>
          <thead>
            <tr>
              <th>Aspect</th>
              <th>Unsalted</th>
              <th>Author-Salted</th>
              <th>Escrowed</th>
            </tr>
          </thead>
          <tbody>
            <tr>
              <td>Who can verify</td>
              <td>Anyone with doc</td>
              <td>Author's choice</td>
              <td>Conditions-based</td>
            </tr>
            <tr>
              <td>Privacy</td>
              <td>None (hash public)</td>
              <td>High (author controls)</td>
              <td>Medium (escrow policy)</td>
            </tr>
            <tr>
              <td>Typical use</td>
              <td>Published works</td>
              <td>Unpublished drafts</td>
              <td>Legal/regulatory</td>
            </tr>
            <tr>
              <td>Salt storage</td>
              <td>N/A</td>
              <td>Author responsibility</td>
              <td>Third party</td>
            </tr>
            <tr>
              <td>Lost salt impact</td>
              <td>N/A</td>
              <td>Cannot verify</td>
              <td>Escrow backup</td>
            </tr>
          </tbody>
        </table>
      </section>
    </section>

    <section anchor="security">
      <name>Security Considerations</name>
      <t>
        This document provides examples for the Proof of Process format.
        Security considerations for the format itself are specified in the
        main architecture document <xref target="I-D.condrey-rats-pop"/>.
      </t>
      <t>
        Example data in this document is illustrative. Implementations
        MUST NOT use example values as actual cryptographic material.
        All hash values, keys, and signatures shown are placeholders.
      </t>
    </section>

    <section anchor="iana">
      <name>IANA Considerations</name>
      <t>
        This document has no IANA actions. All registrations are specified
        in the main architecture document.
      </t>
    </section>
  </middle>

  <back>
    <references>
      <name>References</name>

      <references>
        <name>Normative References</name>

        <reference anchor="RFC2119" target="https://www.rfc-editor.org/info/rfc2119">
          <front>
            <title>Key words for use in RFCs to Indicate Requirement Levels</title>
            <author fullname="S. Bradner" initials="S." surname="Bradner"/>
            <date month="March" year="1997"/>
          </front>
          <seriesInfo name="BCP" value="14"/>
          <seriesInfo name="RFC" value="2119"/>
        </reference>

        <reference anchor="RFC8174" target="https://www.rfc-editor.org/info/rfc8174">
          <front>
            <title>Ambiguity of Uppercase vs Lowercase in RFC 2119 Key Words</title>
            <author fullname="B. Leiba" initials="B." surname="Leiba"/>
            <date month="May" year="2017"/>
          </front>
          <seriesInfo name="BCP" value="14"/>
          <seriesInfo name="RFC" value="8174"/>
        </reference>

        <reference anchor="RFC8949" target="https://www.rfc-editor.org/info/rfc8949">
          <front>
            <title>Concise Binary Object Representation (CBOR)</title>
            <author fullname="C. Bormann" initials="C." surname="Bormann"/>
            <author fullname="P. Hoffman" initials="P." surname="Hoffman"/>
            <date month="December" year="2020"/>
          </front>
          <seriesInfo name="STD" value="94"/>
          <seriesInfo name="RFC" value="8949"/>
        </reference>
      </references>

      <references>
        <name>Informative References</name>

        <reference anchor="I-D.condrey-rats-pop">
          <front>
            <title>Proof of Process: An Evidence Framework for Digital Authorship Attestation</title>
            <author fullname="David Condrey" initials="D." surname="Condrey"/>
            <date/>
          </front>
          <seriesInfo name="Internet-Draft" value="draft-condrey-rats-pop-00"/>
        </reference>
      </references>
    </references>
  </back>
</rfc>
