malcolm.wallace at me.com
Sun Jun 26 13:01:17 CEST 2011
On 26 Jun 2011, at 01:53, Tony Morris wrote:
> Having only had a flirt with Data.Time previously, I assumed
> it would be robust like many other haskell libraries.
If, by lack of robustness, you mean that you get runtime errors, then consider them bugs, and file them with the author/maintainer accordingly.
If you mean something else, then being more specific might be useful. I know that the first time I looked seriously at Data.Time it seemed rather byzantine and labyrinthine. So many types! So few direct conversions between them! But when you think more closely about the domain, you realise that notions of time are not simple at all, and have varied widely over history, and the complexity of Data.Time only reflects the complexity of the domain.
The old-time package is still available, and has a much simplified approach to time (which is evidently wrong in many places), but may better suit the needs of applications that only care to be approximate.
More information about the Haskell-Cafe