[Bioc-devel] Survey results: How and why CRAN and Bioconductor developers handle breaking changes differently

[Chris Bogart]

As promised,here are the results: of the survey about breaking changes
we ran last fall:

   http://breakingapis.org/survey/

The R ecosystem stood out in some interesting ways; for example
package authors in CRAN and Bioconductor are more likely than other
ecosystemsto get personally contacted when a package they depend on
changes:

     http://breakingapis.org/survey/downstream.html#mon_personal_contact?ecos=Node_js_NPM,R_CRAN,R_Bioconductor

...and they’re a little more likely to keep dependencies up to date:

    http://breakingapis.org/survey/downstream.html#upgr_dont_update?ecos=R_CRAN,R_Bioconductor

That’s consistent with some of CRAN and Bioconductors’ repository policies.

One thing we were especially curious about was how people perceived
the barriers to getting package updates out due to the two
communities’ diferent ways of vetting and batching updates.
Developers thought that the Bioconductor community in general values
rapid dissemination of package updates more than CRAN community does:

   http://breakingapis.org/survey/values.html#rapid?ecos=R_CRAN,R_Bioconductor

but we were a little surprised to learn that neither community thinks
this is much of a problem:

   http://breakingapis.org/survey/health.html#not_rapid?ecos=R_CRAN,R_Bioconductor

We’re interested in your impressions of the results.  Do the responses
make sense to you?  What answers would you have expected?  Do you
think the differences are intentional?  I’ll try to keep up with
comments here, or you can also send us comments through the website.

We want to sincerely thank the large number of people in the R
community who responded, and we’re eager to hear what you think!

Chris, Anna, Jim, and Christian