Arbitrary abbreviations in phobos considered ridiculous

[H. S. Teoh]

On Thu, Mar 08, 2012 at 08:48:43PM -0800, Jonathan M Davis wrote:
[...]
> I really don't think that Phobos is really any more quirky or
> inconsistent than your average standard library. It's not perfect, but
> it isn't particularly inconsistent either.

In fact, in spite of its present shortcomings, Phobos as a whole is a
lot *more* consistent than, say, your average C library.


> We'll continue to make improvement to it (primarily by adding new
> stuff), but it's increasingly costly to make breaking changes. And, on
> the whole, it's not like what we have is horrible. The biggest
> problems involve whole modules (which are generally older) which need
> to be redesigned, and those will happen.

Personally, I can't wait for the day std.io merges, and std.stream and
std.stdio are deprecated. The current std.stream/std.stdio dichotomy
just makes no sense at all. Plus, a significant part of std.stdio needs
to be tweaked to use Ranges anyway.


> But minor stuff like tweaking function names doesn't really buy us
> enough to be worth it anymore. If a function changes sufficiently to
> merit a full replacement, then maybe we can change its name and phase
> out the old one (e.g. if we change the functions in std.string which
> take patterns to take regexes instead), but changing a name to change
> a name just isn't worth it when we're trying to provide a serious
> offering with D and Phobos.  We're too far along.
[...]

True.

And +1 for std.string functions to take regexes. In this day and age,
hack jobs for doing string operations just don't cut it anymore. Regexes
rule string processing. (Well, they do have their limitations, but
anything *weaker* than regexes is certainly too weak to be generally
useful.)


T

-- 
Guns don't kill people. Bullets do.