1/ Looking at some of that backlash following the release of @MRC_Outbreak code used to model #COVIDー19 I think some context may be useful. I came to epidemiology coming from a different field (pop genetics) were open-source culture was pretty strong, especially in #RStats
2/ Epidemiology was in a different place. Lots of code used internally, little open-source culture, so that 'good' coding practices designed to make code clear, easy to review and maintain, were not very frequent.
3/ I have been pleased to see things change for the better over the years. A lot. The @MRC_Outbreak has been very supportive of this move towards more open practices, and was instrumental helping @RECONEPI get off the ground.
4/ These changes are mostly cultural, and take time. But a huge majority of modellers I know try to be as open as possible. The amount of information available online is much higher than say during the 2009 flu pandemic.
5/ Any 13+ year-old codebase outside professional software development is quite likely to be messy. I know mine is, and am filled with dread every time there's a bug report for #rstats adegenet, despite its 4000+ citations.
6/ Open & transparent is safer, and better. Great that old codebases are now shared: that means they can be improved through reviews from the community.
