1.0.0 release?

[Datseris]

Hi, what's the plan for a 1.0.0 tag? What is left to be done? In general some stability promise would go a long way for developers like us (JuliaDynamics)

[wildart]

I do not recall any particular plan for versioning. We bump it as the package got updated. As for stability, the API haven't changed for years: fit/transform[/reconstruct].

[Datseris]

I do not recall any particular plan for versioning.

I thought the plan for all accepted Julia packages was to follow this: https://semver.org/ .

As for stability, the API haven't changed for years: fit/transform[/reconstruct].

I do not understand. Then what's the problem with releasing a version 1.0.0?

It is important for people depending here. As instructed by SymVer, you can do whatever change you want to a non-major version. If you release 1.0.0, this would allow us safely to depend on versions [1.0.0, 2.0.0) without ever having a worry.

As a (very tiny) developer of Julia packages, I don't understand your comment. You can still release a 2.0.0 version if in "years" from now you want to do breaking changes. Please help me understand what "problems" are there in releasing a major version.

[wildart]

Please help me understand what "problems" are there in releasing a major version.

There are no "problems" per se but some outstanding issues.

In my opinion, for a proper long lasted release following issues need to be addressed first:

• cMDS API inconsistency (related to Returning eigenvalues from classical_mds? #60)
• some type instabilities (Make CCA take a type parameter #53)
• a ICA convergence issue (Changed way the cost of unmixing matrix calculated it ICA #45).

Of course, documentation and examples need to be straighten up. I like Documenter.jl set of features (e.g. doctests) and a unified look.

[wildart]

All things done for 1.0.0 release. It will be out shortly.