The reports listed below were automatically generated during testing. Some care needs to be taken when interpreting them:
- Percentages shown are for the number of test files that were classified correctly
- This does not necessarily correlate with the "percentage of the spec" a validation method tests
- This certainly does not correlate with the number of errors a method will catch in the wild
- The test set includes things which the spec does not mention, but does not explicitly disallow, i.e. rational numbers or numbers with a non-decimal base in MathML
- Tools may be more strict than the spec, or choose to correct errors, for example by not allowing "123" as an identifier (valid according to the 1.0 spec)
- Tools not specifically aimed at validation will often try to be inclusive - as a feature
- DTD validation (72%)
- Schema validation (72%)
- RelaxNG validation (77%)
- Cellmlmanip (75%)
- Myokit (83%)
- OpenCOR / CellML API (80%)