Available with limitations

Preserve the evidence behind every result.

A result that cannot name its dataset, its plan and its software version is hard to defend. ValidTrials stores checksums on frozen datasets and can keep the method and software version with a job. Re-running a job later should use that snapshot, not a silent newer extract. This is not an external reproducibility certificate.

What this capability does

Snapshots

A freeze has a checksum and an entry count.

Plans

The analysis plan version is stored with the job.

Software versions

The method records the library versions it used.

No silent refresh

A later data change does not rewrite the frozen file.

Exports

CDISC-shaped packages are built from a freeze whose specification hash still matches.

Limits

Independent re-execution on another organization's computer is a process you run. It is not a button that proves identity.

How it works

Freeze

The checksum is stored.

Run

The job names that freeze.

Store

The output and the versions stay together.

Re-read

A reviewer uses the same snapshot.

Connected modules and services

Questions

Do you guarantee bitwise identity forever?

We store the checksum and the versions. A future library change is a reason to keep the old environment, not a promise that today's container will exist unchanged.

Is this an archive in the legal sense?

Retention is your policy. The export is a file set.

Available with limitations means the described software behavior exists. A professional service is work a person does under a separate scope. Nothing here is a certification.

See the workflow on your protocol.

The conversation is about your study design and what the software can and cannot do today.