search

usnistgov / ESV-Server

Entropy Source Validation Protocol and Server specifications
13 stars 12 forks source link

The spec does not appear to be updated for version 1.3.1 #31

Closed mmccarl closed 1 year ago

mmccarl commented 1 year ago

the release notes indicate that alphabetsize has been removed from the registration and all conditioning components require h_out, but the spec doesn't agree with these notes. Are any other changes required?

celic commented 1 year ago

Those should be the only changes to the spec. I'll get this updated.

celic commented 1 year ago

See https://github.com/usnistgov/ESV-Server/pull/32

Also removes ITAR as a registration property.

mmccarl commented 1 year ago

thanks for making the changes...

with respect to 'repeatability' and versioning, would this constitute a new version of the endpoint (or esvVersion)? We have existing JSON's in our database that would may now fail and would therefore not be repeatable if we submitted them.

i bring this up as a general concern of "when does the version number need to change?"

celic commented 1 year ago

That is a good question. I think an auditor (probably me) would need to acknowledge this. I do not want to up the version over minor elements. It would be best practice to know what version of the server was used on a particular request, but that could also be figured out afterwards based on the releases tracked by this repository.

I think the repeatability of submissions to ESVTS is not a high priority within the scope. Repeatability of statistical testing is significantly more important. If we do revise this scope, I think this would be important to reflect in the requirements. Thank you for pointing this out.