Douglas McClean

Perhaps instead some kind of wrapper library around chart, to force you to pick units and then help you tack their names on to the end of your axis labels....

Other possible names include `Measurement`, `Observation`, `WrittenQuantity`, `SerializedQuantity`, `ExpressedQuantity`. I'm not sold on any of these.

I rather like `PresentationQuantity` or `PresentedQuantity`. It doesn't carry the same baggage about who is presenting to whom (it works for both input and output scenarios) and doesn't imply that...

It now occurs to me based on the name `PresentationQuantity` that this is also the way to model #82, I think. The rounding happens during the conversion to a `PresentationQuantity`...

@bjornbm If you have a minute, could you comment on this? In particular things like #8 require taking a dependency on either `parsec` or `attoparsec` but as they don't involve...

Instances would be useful for interoperating with these, possibly among others: `linear`, `vector-space`, `HaTeX`, `ihaskell`.

OK, tentative plan is below: Make 4 projects - `dimensional` - `dimensional-codata` (split off from dimensional-dk-experimental, only outside of dimensional itself because they change it periodically, potentially could be consolidated?)...

Similar flags for `aeson`, `binary`, and any other serialization instances that anyone eventually wants.

The downsides to restructuring and keeping everything in one repo are more obvious once I started to do it. Of note are issues with setting up the CI build(s) and...

Spun off: - https://github.com/dmcclean/dimensional-codata - https://github.com/dmcclean/dimensional-attoparsec Primarily just boilerplate for appeasing git, cabal, and travis. Substantive commits to come, along with hackage release for `dimensional-codata` since it is essentially already...