Assessment Scale
All technologies in the catalog are evaluated along two independent dimensions.
Autonomy Level (A0–A3)
Measures operational independence from the cloud and the internet.
| Level | Description |
|---|---|
| A0 | Cloud-bound – requires constant internet connection and external accounts to function. |
| A1 | Online-dependent – can work offline partially, but still needs occasional external services or authentication. |
| A2 | Offline-capable – works completely without internet; data stored locally, but may lack built-in data export or recovery mechanisms. |
| A3 | Fully autonomous – offline-first, local data storage, full data export, and built-in pause/exit/recoverability. |
Transparency Level (T0–T2)
Measures architectural openness and auditability.
| Level | Description |
|---|---|
| T0 | Closed – no public documentation of internals; source code not available. |
| T1 | Documented – architecture is described in public documents (whitepapers, security guides), but source code may be closed. |
| T2 | Open-source – source code is publicly available under an OSI-approved license; fully auditable. |
Relationship with whose.world criteria
The whose.world framework provides the philosophical foundation for TAS. The full Infrastructure Audit applies eight questions to any technology. They map to the scales as follows:
Three structural questions – Autonomy Level
- Pause – can you stop the system at any moment?
- Exit – can you leave with all your data?
- Recoverability – can you roll back to a previous state?
These three directly determine the A-level. A system where all three are “yes” is A3. A system where all three are “no” is A0.
Visibility – Transparency Level
- Visibility – can you inspect how the system works?
This determines the T-level. Open source with an OSI license = T2. Documented but closed = T1. Opaque = T0.
Five diagnostic questions – beyond A/T
The A/T rating is necessary but not sufficient. Five additional questions reveal forces that the rating alone doesn’t capture:
- Personalisation – does it build a behavioural model of you?
- Urgency – does it manufacture time pressure?
- Hidden cost – what do you pay besides money?
- Transparency fragility – does its value depend on your ignorance?
- Trajectory – is the project moving toward openness or closure?
A tool can be A3/T2 today and still have aggressive telemetry, hidden costs, or a closing trajectory. The A/T rating tells you where the tool stands. The diagnostic questions tell you what’s pulling at it – and where it’s headed.
Together, they form a complete picture of technological sovereignty.
How to score — step by step
Step 1: Determine A level (structural questions)
Answer the three structural questions. Use this table:
| Pause | Exit | Recoverability | A level |
|---|---|---|---|
| ✅ yes | ✅ yes | ✅ yes | A3 |
| ✅ yes | ✅ yes | ⚠️ partial | A2 |
| ✅ yes | ⚠️ partial | ⚠️ partial | A2 |
| ✅ yes | ❌ no | any | A1 |
| ⚠️ partial | any | any | A1 |
| ❌ no | any | any | A0 |
Rule: the weakest answer dominates. If Exit is ❌, the tool is A0 or A1 regardless of the other answers.
Step 2: Determine T level (transparency)
| Source code | License | T level |
|---|---|---|
| Public, OSI-approved license | MIT, Apache, GPL, AGPL, MPL… | T2 |
| Architecture documented, code closed | Whitepaper, security audits published | T1 |
| Closed, no public documentation | Proprietary, no auditable information | T0 |
Note: BSL (Business Source License) is NOT OSI-approved → T1, not T2.
Step 3: Determine D score (diagnostic questions)
Each of the five diagnostic questions is answered ✅ (clean) or ⚠️ (concern):
| Question | ✅ clean | ⚠️ concern |
|---|---|---|
| Personalisation | No behavioural profiling | Builds user model, telemetry |
| Urgency | No artificial pressure | Countdown timers, forced upgrades |
| Hidden cost | No non-monetary costs | Privacy, attention, lock-in |
| Transparency fragility | Value survives full visibility | Value depends on user ignorance |
| Trajectory | Moving toward openness | License changes, feature gating, cloud push |
Count the ⚠️ concerns:
| ⚠️ concerns | D score | Trajectory |
|---|---|---|
| 0 | D5 | opening |
| 1 | D4 | stable |
| 2 | D3 | mixed |
| 3 | D2 | closing |
| 4–5 | D1 | closing |
Step 4: Write the TAS Score
Format: S_/3 — D_/5 · A_ / T_
Example: S3/3 — D4/5 · A3 / T2
- S = structural score = number of ✅ among the 3 structural questions
- D = diagnostic score = 5 minus number of ⚠️ concerns
- A = autonomy level from Step 1
- T = transparency level from Step 2
If S=3 and D=5 — no explanation needed. If S<3 or D<5 — always explain which question caused the deduction.
Worked example: PhotoPrism
Structural:
- Pause ✅ — can be stopped cleanly
- Exit ✅ — original files remain untouched
- Recoverability ✅ — standard backup applies
→ S3/3 · A3
Transparency:
- AGPL-3.0, fully open source
→ T2
Diagnostic:
- Personalisation ✅ — no behavioural profiling
- Urgency ✅ — no artificial pressure
- Hidden cost ✅ — no non-monetary costs
- Transparency fragility ✅ — value survives full visibility
- Trajectory ⚠️ — PhotoPrism+ is a paid closed tier; some features gated
→ D4/5 · mixed
Final: S3/3 — D4/5 · A3 / T2 · trajectory: mixed