[Neuroimaging] Planning for data formats - upcoming journal club

[Matthew Brett]

Hi,

On Fri, Nov 12, 2021 at 2:45 PM Joshua Moore (Staff)
<j.a.moore at dundee.ac.uk> wrote:
>
> > Thanks - these are strikingly relevant ...
>
>
>
> I’d second that! But I (with my OME hat on) would also second Satra said that there are cases & users that will continue to prefer HDF5 and once we’ve worked through the design of NGFF, we’ll likely try to bring the same design to HDF5, keeping the two convertible to one another to support as much of the community as possible.
>
>
>
> And for what it’s worth, we also received CZI funding (with my Zarr hat on) for driving imaging forward, and the kind folks at Unidata are doing what they can to have HDF5 & Zarr supported seamlessly through NetCDF. Exciting times.

Thanks - yes - it would be very good if you had time to come to the
journal club - Doodle:

https://doodle.com/poll/5vwkdt7zic2bu3wn

I think we're going to be reflecting there on:

* Whether we should think in terms of API (e.g. Zarr) rather than data format
* Whether we should distinguish between a sharing format and a
for-processing format.
* What criteria we should be applying to each of these formats.

I'm sure you've thought hard about these, and it would be very helpful
to have your thoughts.   I would also, personally, love to see a
show-and-tell for the various parts we are thinking about, such as
Zarr, HDF5 / Minc, NeuroJSON (http://neurojson.org), and so on.  And
maybe "Why I will never use an HDF5 / NeuroJSON / <your format here>
format. Perhaps we could plan that after the journal club ...

Cheers,

Matthew