CRAN release

[seabbs]

We (mostly thanks to @sbfnk) are starting to accumulate a lot of nice new features all (I think) are well implemented enough not to break backwards compatibility. Given that we should likely plan for a CRAN release in the near future. Especially as some of the larger changes may have to lead to breaking changes.

[sbfnk]

Question is whether we do this now or wait until the 2.0.0 release (#423) is ready. Given quite substantial further development that will be needed especially to update the observation model and covariate interface I agree that there might be a case for an intermediate release as long as we're fairly confident that features have stabilised (avoiding repeating the introduction/deprecation of dist_spec necessitated by the {rstan} update).

[seabbs]

Yeah I think its basically as we at a point with few breaking changes but new features that people could use before things get dangerous?

I think we probably are.

[sbfnk]

Yes, I agree. Probably worth just waiting until we're happy the distribution interface and then going through some of the minor issues for release?

[seabbs]

Agree. Its at essentially the same point as #538 can happen imo

[sbfnk]

There now is a project board at https://github.com/orgs/epiforecasts/projects/9 with the fairly minor issues that might be nice to tackle for this.

[sbfnk]

There are still quite a few updates in the 2.0.0 plan. It might be nice to plan for a few intermediate releases which brings in each of these separately, followed by a "clean" 2.0.0 release where all deprecated functions are removed.

[sbfnk]

Tentatively:

• 1.5.0: new distribution interface Reconsider distribution interface. #382
• 1.6.0: extend distribution interface to continuous parameters Change all priors to new dist interface, not just delays #525
• 1.7.0: covariate interface switch to directly calculated growth rate and extend model options, take 2 #345
• 1.8.0: combine estimate_infections and estimate_secondary stan models into one fold secondary options into observation model #601
• 1.9.0: support multiple observation time series Fit to multiple time series and fold in secondary model #371
• 2.0.0: clean version with deprecated functionality removed

Other minor issues can slot in where they fit in and major ones left to later releases.