[yt-dev] Towards a 3.0 release

[Matthew Turk]

 Hi Anthony,

Let me start out by saying, I have a tendency to put off this type of
thing, so thank you for bringing it up.

On Mon, Sep 23, 2013 at 12:56 AM, Anthony Scopatz <scopatz at gmail.com> wrote:
> Hey yt,
>
> I realize I am being fairly selfish here, but I think that it
> would be great if there was a stable yt-3.0 that was the
> yt.  My personal opinion is that it is stable enough.

I agree with this.

>
> I don't expect this to happen tomorrow.  But I don't think
> that the end of the year would be an unreasonable target.
>
> Matt seemed amicable to the idea when we talked about it
> yesterday.  Still, I thought I would throw this out there
> for folks to give feedback on.

Britton's point is good, that not everything has been ported; there
are a few major things that I can think of.  This includes clump
finding, several of the frontends, and boolean objects.  I think I
know how to do all of these, they just need time.

And, there is the bigger problem of having created a list of features
that 3.0 was to have and not getting all of them implemented.  Units
is the biggest one, but that can be added later.  These can basically
be divided into a few categories:

 * Things that break backwards compat and so need to be implemented
before a "release"
 * Things we'd like but that don't break backwards compat and so can
be added later

That all being said, "perfect is the enemy of good enough."  Since the
2.X series will be a long-term support branch, I think if we can
finish up ripping bandaids off, we should try for a 3.0 release by the
end of the year.  That puts the following things on the critical path:

 * Documentation, which means both a cheat sheet of what is new as
well as a bunch of explanations of what has changed under the hood.
Much of this will fall to me.
 * Porting remaining frontends to 3.0, which unfortunately also means
pushing hard on the patch-IO refactor.
 * Field renaming, which currently is caught up in the units refactor.
 Units likely won't be done until 3.1 or 3.11 (yt for workgroups) so
we may need to cherry pick these changes out.  (I worked on units last
week, though, and it may not be as far off as that.)

There are a few YTEPs as well that *need* to be implemented and tested:

https://bitbucket.org/yt_analysis/ytep/pull-request/25/ytep-0017-de-astroifying-yt-a-bit-and
https://bitbucket.org/yt_analysis/ytep/pull-request/24/ytep-0018-reducing-dict-like-access-to

So if we summarize these, it comes down to:

 1. Ripping off bandaids for API changes.
 2. Documenting.

I think a 3.0 by the new year is feasible, but only if we can finish
both of those, and only if we allow that not *everything* will be
finished by then.

-Matt

>
> Be Well
> Anthony
>
> _______________________________________________
> yt-dev mailing list
> yt-dev at lists.spacepope.org
> http://lists.spacepope.org/listinfo.cgi/yt-dev-spacepope.org
>