Skip to Content

Changes for page Workflows

Last modified by Robert Schaub on 2026/02/08 08:25
From version 1.2
edited by Robert Schaub
on 2025/12/23 18:01
Change comment: Update document after refactoring.
To version 1.3
edited by Robert Schaub
on 2026/01/20 20:25
Change comment: Renamed back-links.

Summary

Details

Page properties
Content
... ... @@ -1,11 +1,15 @@
1 1  = Workflows =
2 +
2 2  FactHarbor workflows are **simple, automated, focused on continuous improvement**.
4 +
3 3  == 1. Core Principles ==
6 +
4 4  * **Automated by default**: AI processes everything
5 5  * **Publish immediately**: No centralized approval (removed in V0.9.50)
6 6  * **Quality through monitoring**: Not gatekeeping
7 7  * **Fix systems, not data**: Errors trigger improvements
8 8  * **Human-in-loop**: Only for edge cases and abuse
12 +
9 9  == 2. Claim Submission Workflow ==
10 10  
11 11  === 2.1 Claim Extraction ===
... ... @@ -13,11 +13,13 @@
13 13  When users submit content (text, articles, web pages), FactHarbor first extracts individual verifiable claims:
14 14  
15 15  **Input Types:**
20 +
16 16  * Single claim: "The Earth is flat"
17 17  * Text with multiple claims: "Climate change is accelerating. Sea levels rose 3mm in 2023. Arctic ice decreased 13% annually."
18 18  * URLs: Web pages analyzed for factual claims
19 19  
20 20  **Extraction Process:**
26 +
21 21  * LLM analyzes submitted content
22 22  * Identifies distinct, verifiable factual claims
23 23  * Separates claims from opinions, questions, or commentary
... ... @@ -24,17 +24,18 @@
24 24  * Each claim becomes independent for processing
25 25  
26 26  **Output:**
33 +
27 27  * List of claims with context
28 28  * Each claim assigned unique ID
29 29  * Original context preserved for reference
30 30  
31 31  This extraction ensures:
39 +
32 32  * Each claim receives focused analysis
33 33  * Multiple claims in one submission are all processed
34 34  * Claims are properly isolated for independent verification
35 35  * Context is preserved for accurate interpretation
36 36  
37 -
38 38  ```
39 39  User submits → Duplicate detection → Categorization → Processing queue → User receives ID
40 40  ```
... ... @@ -46,6 +46,7 @@
46 46  {{include reference="FactHarbor.Specification.Diagrams.Claim and Scenario Workflow.WebHome"/}}
47 47  
48 48  == 3. Automated Analysis Workflow ==
56 +
49 49  ```
50 50  Claim from queue
51 51  ↓
... ... @@ -65,18 +65,25 @@
65 65  ```
66 66  **Timeline**: 10-30 seconds
67 67  **90%+ published automatically**
76 +
68 68  == 3.5 Evidence and Verdict Workflow ==
78 +
69 69  {{include reference="FactHarbor.Specification.Diagrams.Evidence and Verdict Workflow.WebHome"/}}
80 +
70 70  == 4. Publication Workflow ==
82 +
71 71  **Standard (90%+)**: Pass quality gates → Publish immediately with confidence scores
72 72  **High Risk (<10%)**: Risk > 80% → Moderator review
73 73  **Low Quality**: Confidence < 40% → Improvement queue → Re-process
86 +
74 74  == 5. User Contribution Workflow ==
88 +
75 75  ```
76 76  Contributor edits → System validates → Applied immediately → Logged → Reputation earned
77 77  ```
78 78  **No approval required** (Wikipedia model)
79 79  **New contributors** (<50 reputation): Limited to minor edits
94 +
80 80  == 5.5 Quality and Audit Workflow ==
81 81  
82 82  {{include reference="FactHarbor.Specification.Diagrams.Quality and Audit Workflow.WebHome"/}}
... ... @@ -83,16 +83,21 @@
83 83  
84 84  
85 85  == 6. Flagging Workflow ==
101 +
86 86  ```
87 87  User flags issue → Categorize (abuse/quality) → Automated or manual resolution
88 88  ```
89 89  **Quality issues**: Add to improvement queue → System fix → Auto re-process
90 90  **Abuse**: Moderator review → Action taken
107 +
91 91  == 7. Moderation Workflow ==
109 +
92 92  **Automated pre-moderation**: 95% published automatically
93 93  **Moderator queue**: Only high-risk or flagged content
94 94  **Appeal process**: Different moderator → Governing Team if needed
113 +
95 95  == 8. System Improvement Workflow ==
115 +
96 96  **Weekly cycle**:
97 97  ```
98 98  Monday: Review error patterns
... ... @@ -109,18 +109,26 @@
109 109  ```
110 110  New algorithm → Split traffic (90% control, 10% test) → Run 1 week → Compare metrics → Deploy if better
111 111  ```
132 +
112 112  == 9. Quality Monitoring Workflow ==
134 +
113 113  **Continuous**: Every hour calculate metrics, detect anomalies
114 114  **Daily**: Update source track records, aggregate error patterns
115 115  **Weekly**: System improvement cycle, performance review
138 +
116 116  == 10. Source Track Record Workflow ==
140 +
117 117  **Initial score**: New source starts at 50 (neutral)
118 118  **Daily updates**: Calculate accuracy, correction frequency, update score
119 119  **Continuous**: All claims using source recalculated when score changes
144 +
120 120  == 11. Re-Processing Workflow ==
146 +
121 121  **Triggers**: System improvement deployed, source score updated, new evidence, error fixed
122 122  **Process**: Identify affected claims → Re-run AKEL → Compare → Update if better → Log change
149 +
123 123  == 12. Related Pages ==
151 +
124 124  * [[Requirements>>FactHarbor.Specification.Requirements.WebHome]]
125 -* [[Architecture>>FactHarbor.Specification.Architecture.WebHome]]
153 +* [[Architecture>>Archive.FactHarbor.Specification.Architecture.WebHome]]
126 126  * [[Data Model>>FactHarbor.Specification.Data Model.WebHome]]