datetime review part 2 [Update 4]

[Jonathan M Davis]

On Tuesday, November 09, 2010 17:34:15 Jonathan M Davis wrote:
> On Tuesday, November 09, 2010 16:23:56 Yao G. wrote:
> > On Tue, 09 Nov 2010 18:11:59 -0600, Jonathan M Davis
> > <jmdavisProg at gmx.com>
> > 
> > wrote:
> > > I think that the real question here is how good the API is.
> > 
> > I think that the sheer size of the library, specially with a datetime.d
> > file with almost 31,600 lines of code (granted, most of them are unit
> > test and documentation), makes a little difficult to analyze or give a
> > proper review of the code. It will takes a fair share time to do it.
> > Maybe that's why very few people has given criticism or suggestions. I
> > only gave a cursory view (you need to scroll a lot just to find the
> > implementation code), but I'll give it a more throughly review tonight.
> 
> That's quite understandable, but that's part of the reason that the ddoc
> html files are there. You don't actually have to go trolling through the
> code to look at the API.

I should probably add that it would be worth it for folks to just look at the 
documentation for what they would generally try and do with a datetime module 
and see whether they can do it reasonably well and what problems they'd have. 
While looking at the module as a whole is definitely needed, just having folks 
look at the portions that they'd use and commenting on that would be useful. Is 
getting the time easy enough and intuitive enough? Is doing what you need to do 
with time in basic applications once you have it easy or intuitive enough? Etc.

- Jonathan M Davis