feedback on pkg

[sckott]

brad emailed me about this, so here's a few thoughts, take em or leave em

• users are likely to have names as a factor if they forget about stringsasfactors, TNRS fails if you pass in a vector of factors
• some tests would be a good idea
• some checking of inputs would be helpful or better API error message, e.g.,

TNRS(taxonomic_names=mtcars)
Error: Bad Request

Either on the R side check that the data.frame provided has the correct stuff, or maybe Brad could send back an informative error message about what went wrong. maybe he does and it's not transmitted

In another eg, you can pass in an empty data.frame, and get back the same error above. Could do a simple check on the R side for that.

• Similar to above for the taxonomic_names parameter, the other params I can pass non-sense to and get the same Bad Request error. Could easily check that a user passed in one of the allowed values on the R side to avoid non-meaningful errors
• an ellipsis ... to pass in curl options would be good so folks can get verbose curl output
• Seems important to point out that results can have more than 1 row of results for each name submitted so users don't assume there's one row per name submitted
• some informaiton or links in the docs about what the returned fields mean would be useful