So, per @ChrisBarker-NOAA's comment on #139, I like the idea of stating that char attributes are constrained to ASCII (latin-1?), and that string attributes should be treated as utf-8. There's always the possibility of adding an encoding attribute at some later date if there is demand.
As much as I like @Dave-Allured's suggestion above, I think it's probably best to leave string array attributes to CF 2.0 - or at least until a later date. It's a pretty pervasive change. It's not hard from a technical standpoint (and my organizational brain loves the idea!), but I think it will be confusing to a number of 'less technical' scientists who I encounter that find netCDF and CF terribly confusing already. There's also quite a few questions that would need to be resolved about **`cell_methods`** and other attributes like it would be affected. Thoughts? -- You are receiving this because you commented. Reply to this email directly or view it on GitHub: https://github.com/cf-convention/cf-conventions/issues/141#issuecomment-433403324
