Astropy 0.1.0 milestone!

[Thomas Robitaille]

Hi everyone,

As described in my previous email, we've made great progress with the Astropy package, and there are a number of other components that should be finished very soon. We (the coordination committee) feel that it is now time to prepare things for a first release - 0.1.0 - so to get things moving, we've decided to set a strict deadline of June 1st for any new functionality to be included in 0.1.0. After this date, we will take a week or so to ensure that things are stable and fix any bugs if needed, before releasing our first stable version by mid-June at the latest.

Therefore, if you are currently working on components that you want to see integrated into the core package for 0.1.0, please make a push to get this finished as soon as possible to ensure that we have enough time to iterate on it if needed. We should have all relevant pull requests and issues closed by June 1st!

Also note that not *everything* has to go in 0.1.0. These are still very early days for Astropy, and what is currently implemented is only a small fraction of what we want the package to become. Don't think of this milestone as the end, it is just the beginning! There will be many opportunities to get involved even if you haven't had a chance yet :-)

If you want to help, but aren't implementing any functionality for 0.1.0, even just installing and using the current developer version and reporting any issues would be very useful!

Cheers,

Tom

[Erik Bray]

Just out of curiosity, was there some triaging of issues for the 0.1.0
milestone? I was just looking through the list of open issues, and
there are some things that I *thought* were slated for 0.1.0, but that
are not there anymore. I could be misremembering though.

Also, are we really going to have things like "coords" settled on in
time for a June release?

I'm all for making a release of some form or another sooner rather
than later, even if it doesn't have everything. Just wondering what
will be culled.

Erik B.

[Thomas Robitaille]

Hi Erik,

> Just out of curiosity, was there some triaging of issues for the 0.1.0
> milestone?  I was just looking through the list of open issues, and
> there are some things that I *thought* were slated for 0.1.0, but that
> are not there anymore.  I could be misremembering though.

I haven't un-tagged anything for 0.1.0 - though we have closed some
tickets that were already implemented. Are there specific things you
were thinking of that don't appear there?

> Also, are we really going to have things like "coords" settled on in
> time for a June release?

We will see - Erik T. mentioned he may have a pull request done for
this in time before the June 1st deadline.

> I'm all for making a release of some form or another sooner rather
> than later, even if it doesn't have everything.  Just wondering what
> will be culled.

Things like photometry and spectral utilities are being worked on in
an affiliated package, so it's unlikely they will be included in the
core in time. I personally feel that whatever we already have
currently is already enough to warrant a first release, since it
already unifies four major python modules into one! I'm hoping we can
have a pretty frequent release cycle, so we can add new functionality
bit by bit (i.e. if certain components miss 0.1.0, they won't have to
wait a year to be released).

Cheers,
Tom

[Wolfgang Kerzendorf]

On 2012-05-10, at 4:11 PM, Thomas Robitaille wrote:

> Hi Erik,
>
>> Just out of curiosity, was there some triaging of issues for the 0.1.0
>> milestone? I was just looking through the list of open issues, and
>> there are some things that I *thought* were slated for 0.1.0, but that
>> are not there anymore. I could be misremembering though.
>
> I haven't un-tagged anything for 0.1.0 - though we have closed some
> tickets that were already implemented. Are there specific things you
> were thinking of that don't appear there?
>
>> Also, are we really going to have things like "coords" settled on in
>> time for a June release?
>
> We will see - Erik T. mentioned he may have a pull request done for
> this in time before the June 1st deadline.
>

As we've seen from the time implementation that even if we have a pull request there's still a lot of (important) discussion, which will take time.
I think we should not rush this and leave it for 0.2. I don't know if we can get the errors/ flags/ masks (in nddata) done in time if we concentrate on this.

>> I'm all for making a release of some form or another sooner rather
>> than later, even if it doesn't have everything. Just wondering what
>> will be culled.
>
> Things like photometry and spectral utilities are being worked on in
> an affiliated package, so it's unlikely they will be included in the
> core in time. I personally feel that whatever we already have
> currently is already enough to warrant a first release, since it
> already unifies four major python modules into one! I'm hoping we can
> have a pretty frequent release cycle, so we can add new functionality
> bit by bit (i.e. if certain components miss 0.1.0, they won't have to
> wait a year to be released).

I fully agree, that none of these will take a year and once we have coord/time implemented that would warrant another release.

> Cheers,
> Tom