docs: document implementation matrix

[codeboten]

The idea w/ this matrix is to provide end users a clear picture of what is implemented for each language implementation. Ideally, the matrix will be:

1. auto generated from the kitchen-sink
2. automatically filled in for every implementation via a test

In the short term, the automation for fill in the implementation details could be manual, but ideally that would be short-lived for the sake of users and implementors.

[jack-berg]

Hey @codeboten - I put together a branch with a little demo that shows how the code generation might work: https://github.com/jack-berg/opentelemetry-configuration/tree/generate-compliance-matrix

It wrote a fast / dirty script to:

1. Read in all the JSON schema files from ./schema
2. Resolve all the $refs
3. Iterate through all the types and generate markdown

See an example of the output of markdown generation here: https://github.com/jack-berg/opentelemetry-configuration/blob/generate-compliance-matrix/schema.md

[marcalff]

@jack-berg Could you check your demo branch ? I can't see the script or the generated markdown. Thanks.

[jack-berg]

Oops looks like I had some copy paste errors on the links. Fixed.

This still needs a decent amount of work before its ready for PR, but I think its conceptually right.

[marcalff]

Oops looks like I had some copy paste errors on the links. Fixed.

This still needs a decent amount of work before its ready for PR, but I think its conceptually right.

Thanks, I can see it now.

Some suggestion below

Instead of presenting:

• a table with "Property, Type, Description, C++, C#, erlang, etc"

consider presenting:

• a table with "Property, Type, Description"
• a table with "Property, C++, C#, erlang, etc"

Also, extension points are missing.

In Samplers for example, I expect an entry for extension points, and the possibility for each SIG to document if extension points are suported or not.

[marcalff]

@codeboten @jack-berg

Any news on this ?

opentelemetry-cpp is about to complete the declarative configuration implementation, and I really would like a place to report details, node by node, in the matrix.

This will help to keep track of changes to the model going forward, at a very detailed level, like for example diffs in modeling between 1.0-rc.1 and 1.0-rc.2.

Also, the specifications now use some tooling with a yaml file per SIG, could the same be used here ?

This will avoid merge collisions between SIGs reporting implementation completeness, and improve the git history (git blame).

[jack-berg]

@marcalff I'll work on refining some ideas sometime in the next 7-10 days.

[jack-berg]

See #312