[ome-devel] Proposed tweak to µManager data files in 2.0

[Nico Stuurman]

On 4/14/16 12:12 PM, Chris Weisiger wrote:
> Hi Hadrien,
>
> On Thu, Apr 14, 2016 at 11:25 AM, Hadrien Mary <hadrien.mary at gmail.com 
> <mailto:hadrien.mary at gmail.com>> wrote:
>
>     Hello guys,
>
>     Sorry for the very naive question Chris but why can't you just use
>     the OME XML format as the core format for MM instead of creating
>     another format ?
>
>     I don't understand the disadvantages of doing that and if there is
>     limitations with the current OME XML format why not trying to
>     improve it with the OME folks ?
>
>
> I'm not foreclosing on any options; quite simply we just *don't have* 
> any plans one way or another right now. If after we do our research we 
> determine that the OME TIFF format (possibly with some enhancements) 
> is the best option, then we'll use it. It certainly has numerous 
> advantages in terms of current user base, not to mention avoiding 
> creating yet another file format. :)

Karl Bellve summarized the main problem for such an approach quite 
nicely in a post to the Micro-Manager mailing list (same thread):

"How you save metadata doesn't matter much when µManager 1.4 (don't know 
about 2.0) doesn't do a very good job of exporting enough types of 
metadata  (look at Zeiss ->OME in comparison). This is due to being 
device centric, and each device determines how it wants to label 
information. We should think about label consistency with camera, 
filters, shutters, etc.  It will make OMEs job easier and other software 
much easier."

Saying it in another way: Micro-Manager devices adapters have a lot of 
freedom to decide how to name things.  Moving towards using pure OME xml 
dictates loss of that freedom and a huge amount of work to rewrite 
device adapters.  This is independent of performance issues writing OME 
xml that would also need to be addressed.

Best,

Nico


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openmicroscopy.org.uk/pipermail/ome-devel/attachments/20160414/b2b5515b/attachment.html>