Skip to Content

Wiki source code of Workflows

Last modified by Robert Schaub on 2025/12/24 18:27
Show last authors
1 = Workflows =
2 FactHarbor workflows are **simple, automated, focused on continuous improvement**.
3 == 1. Core Principles ==
4 * **Automated by default**: AI processes everything
5 * **Publish immediately**: No centralized approval (removed in V0.9.50)
6 * **Quality through monitoring**: Not gatekeeping
7 * **Fix systems, not data**: Errors trigger improvements
8 * **Human-in-loop**: Only for edge cases and abuse
9 == 2. Claim Submission Workflow ==
10
11 === 2.1 Claim Extraction ===
12
13 When users submit content (text, articles, web pages), FactHarbor first extracts individual verifiable claims:
14
15 **Input Types:**
16 * Single claim: "The Earth is flat"
17 * Text with multiple claims: "Climate change is accelerating. Sea levels rose 3mm in 2023. Arctic ice decreased 13% annually."
18 * URLs: Web pages analyzed for factual claims
19
20 **Extraction Process:**
21 * LLM analyzes submitted content
22 * Identifies distinct, verifiable factual claims
23 * Separates claims from opinions, questions, or commentary
24 * Each claim becomes independent for processing
25
26 **Output:**
27 * List of claims with context
28 * Each claim assigned unique ID
29 * Original context preserved for reference
30
31 This extraction ensures:
32 * Each claim receives focused analysis
33 * Multiple claims in one submission are all processed
34 * Claims are properly isolated for independent verification
35 * Context is preserved for accurate interpretation
36
37
38 ```
39 User submits → Duplicate detection → Categorization → Processing queue → User receives ID
40 ```
41 **Timeline**: Seconds
42 **No approval needed**
43
44 == 2.5 Claim and Scenario Workflow ==
45
46 {{include reference="FactHarbor.Specification.Diagrams.Claim and Scenario Workflow.WebHome"/}}
47
48 == 3. Automated Analysis Workflow ==
49 ```
50 Claim from queue
51 ↓
52 Evidence gathering (AKEL)
53 ↓
54 Source evaluation (track record check)
55 ↓
56 Scenario generation
57 ↓
58 Verdict synthesis
59 ↓
60 Risk assessment
61 ↓
62 Quality gates (confidence > 40%? risk < 80%?)
63 ↓
64 Publish OR Flag for improvement
65 ```
66 **Timeline**: 10-30 seconds
67 **90%+ published automatically**
68 == 3.5 Evidence and Verdict Workflow ==
69 {{include reference="FactHarbor.Specification.Diagrams.Evidence and Verdict Workflow.WebHome"/}}
70 == 4. Publication Workflow ==
71 **Standard (90%+)**: Pass quality gates → Publish immediately with confidence scores
72 **High Risk (<10%)**: Risk > 80% → Moderator review
73 **Low Quality**: Confidence < 40% → Improvement queue → Re-process
74 == 5. User Contribution Workflow ==
75 ```
76 Contributor edits → System validates → Applied immediately → Logged → Reputation earned
77 ```
78 **No approval required** (Wikipedia model)
79 **New contributors** (<50 reputation): Limited to minor edits
80 == 5.5 Quality and Audit Workflow ==
81
82 {{include reference="FactHarbor.Specification.Diagrams.Quality and Audit Workflow.WebHome"/}}
83
84
85 == 6. Flagging Workflow ==
86 ```
87 User flags issue → Categorize (abuse/quality) → Automated or manual resolution
88 ```
89 **Quality issues**: Add to improvement queue → System fix → Auto re-process
90 **Abuse**: Moderator review → Action taken
91 == 7. Moderation Workflow ==
92 **Automated pre-moderation**: 95% published automatically
93 **Moderator queue**: Only high-risk or flagged content
94 **Appeal process**: Different moderator → Governing Team if needed
95 == 8. System Improvement Workflow ==
96 **Weekly cycle**:
97 ```
98 Monday: Review error patterns
99 Tuesday-Wednesday: Develop fixes
100 Thursday: Test improvements
101 Friday: Deploy & re-process
102 Weekend: Monitor metrics
103 ```
104 **Error capture**:
105 ```
106 Error detected → Categorize → Root cause → Improvement queue → Pattern analysis
107 ```
108 **A/B Testing**:
109 ```
110 New algorithm → Split traffic (90% control, 10% test) → Run 1 week → Compare metrics → Deploy if better
111 ```
112 == 9. Quality Monitoring Workflow ==
113 **Continuous**: Every hour calculate metrics, detect anomalies
114 **Daily**: Update source track records, aggregate error patterns
115 **Weekly**: System improvement cycle, performance review
116 == 10. Source Track Record Workflow ==
117 **Initial score**: New source starts at 50 (neutral)
118 **Daily updates**: Calculate accuracy, correction frequency, update score
119 **Continuous**: All claims using source recalculated when score changes
120 == 11. Re-Processing Workflow ==
121 **Triggers**: System improvement deployed, source score updated, new evidence, error fixed
122 **Process**: Identify affected claims → Re-run AKEL → Compare → Update if better → Log change
123 == 12. Related Pages ==
124 * [[Requirements>>FactHarbor.Specification.Requirements.WebHome]]
125 * [[Architecture>>FactHarbor.Specification.Architecture.WebHome]]
126 * [[Data Model>>FactHarbor.Specification.Data Model.WebHome]]