Rethink canned examples

[lindsayplatt]

• It was kind of confusing to keep flipping between trainR and demoPackage
• Though both seemed necessary
• Maybe we need to add more to trainR, so that once we get through that initial example of how to locally set up a package w/ demoPackage we should not revisit

Suggestions/discussion:
Don’t have trainR setup so much before hand. Have people take on initial issues (create X function) and as the last part of that issue, create the next issue for that function. Start with a couple fully functional parts (function, documentation, tests, checks passing) where NAMESPACE is already setup, and they add on (helps eliminate some of those initial merge conflict issues). Everyone’s issues are enhancements. Have students peer review each others’ enhancement functionality. Have branch w/ functions that have bugs in them and merge in for that section (but shouldn’t impact what’s already been done). Good function(s) could have errors in them already - shouldn’t impact build. Could also have separate errant package for the debugging section. If people are a group (like wateRuse), would we use any of these things? Could have Emily’s proj mgmt spiel beforehand so people have already thought about goals for their pkg. Force people to use trainR or have their own thing to work on. Other USGS package issues instead of trainR issues - free help, real contributions, but could be complex (good for debugging).