why the lucky stiff [mailto:yaml-core@...] wrote:
> We're trying to limit the number of builtin types at the
> moment. Just last month no one was quite happy with the
> timestamp, so we had discussed expansion, even to encompass
> durations. Then we started to lean toward removal of the
> timestamp. Yeah.. slippery slope.. _exactly_.
To clarify: there aren't really "built-in" types in the sense of "types that
any YAML document is required to use". Except for map/seq and str, that is.
Anything else is up to the specific document schema.
That said, we do want all our "recommended" types to be able to work
together and offer a coherent, useful typing framework for schemas to use.
Time is as tricky to get right as currency, and the simple timestamp type we
have is obviously just a partial solution. If we want our "recommended
types" gallery to be comprehensive, we'll have to tackle this eventually.
> So instead of adding more builtin types, we want to focus on
> a useful schema. For adding both explicit and implicit
> types. We've knocked around some ideas. I think we're all
> anxious to see more usage of YAML, more implementations,
> more working code before we tromp that way.
Get latest updates about Open Source Projects, Conferences and News.