On 15/01/2015 7:34 p.m., NVolcz wrote:
On Thursday, 15 January 2015 at 03:42:39 UTC, Rikki Cattermole wrote:
Personally I believe we should move towards a model where every
library is represented by JSON (maybe with help of e.g. markdown) that
gets e.g. parsed by client side technologies such as json.
Benefits:
- We can have central documentation as part of dub registry for all
libraries registered
- Same look and feel for every piece of documentation
- Less dependencies between druntime/phobos docs and dlang.org
Negatives:
- Loss of control over end formatting excluding what e.g. markdown
support
- If javascript is disabled, nothing gets displayed
Nice work!
Central searchable documentation for third party libraries registered on
code.dlang.org could be a killer :-).
Thanks, and I agree it could very well be.
But it will be quite a lot of work, even if everyone agrees this is the
right direction to go.
Eventually this could allow for making e.g. std.algorithm as a subpackage.
But atleast we are getting round to sorting out our documentation now.
