Stating what a data source cannot do is more useful to users than any claim about what it can.
Each of the limits below is deliberate, and each exists because the alternative causes harm.
It does not assess quality
- Records that a work exists, not whether it is good.
- Quality assessment requires reading the work — no data source substitutes for that.
- Metrics recorded are context, never judgement.
- A registry that assessed quality could not be corrected when wrong.
It does not rank
- The underlying data does not support ranking.
- Coverage differs between the things that would be ranked.
- Rankings invite exactly the misuse the data cannot bear.
- Aggregates are provided; ordering them is not.
It does not investigate
- Records formal outcomes — retractions, corrections — with sources.
- Never records allegations, suspicions or open concerns.
- Investigation belongs to bodies with jurisdiction and due process.
- Publishing a suspicion harms a person in ways a later correction does not undo.
It is not complete
- Local-language and applied work is systematically underrepresented.
- Unpublished contributions leave no trace.
- Coverage varies by field, language and venue type.
- A sparse record indicates coverage, not activity — and this must be stated wherever records are shown.
It does not replace judgement
- Not suitable for individual assessment.
- Not suitable for any decision the person cannot see and contest.
- Provides context for judgement, not a substitute for it.
- Users will assume otherwise unless told — which is why it must be documented.
What it does instead
- Records verifiable facts with their sources.
- Maintains stable identity over time.
- Preserves history rather than overwriting it.
- Accepts and propagates corrections.
- States its own coverage and limitations.
One thing worth remembering
A data source that states its limits clearly is more useful than one that implies it has none.
The second is not more capable — it is simply leaving its users to discover the limits themselves, usually after they have relied on it.
Câu hỏi thường gặp
Does a registry assess research quality?
No — it records that a work exists, not whether it is good, since quality assessment requires reading the work and a registry that assessed it could not be corrected when wrong.
Why does it not rank?
Because the underlying data does not support ranking, coverage differs between the things ranked, and rankings invite exactly the misuse the data cannot bear.
Does it investigate misconduct?
No — it records formal outcomes with sources but never allegations or open concerns, since investigation belongs to bodies with jurisdiction and due process.
Is the record complete?
No — local-language and applied work is systematically underrepresented, unpublished contributions leave no trace, and a sparse record indicates coverage rather than activity.
Why state limits explicitly?
Because a source implying it has none is not more capable — it simply leaves users to discover the limits themselves, usually after relying on it.
