[ome-users] Fwd: Problems and suggestions about OMERO

[Aleksandra Tarkowska]

Forwarding to the list.

> From: Olivier Cinquin <ocinquin at uci.edu>
> Date: 2 January 2010 05:53:47 GMT
> To: comments at openmicroscopy.org.uk
> Subject: Problems and suggestions about OMERO
>
> Hi,
> I am considering having my lab use OMERO to store all of our  
> microscopy data. I have installed the server and played with the  
> client, and I have run into problems I'd like to report. I also have  
> suggestions for improvements.
>
> The three major problems I found are the following:
> - TIFF images created in ImageJ by the Bio-Formats Importer plugin  
> from LSM files lose most of their metadata when imported into OMERO.  
> In particular, stacks lose the z-interval, which is a major problem  
> for us. Everything that shows up in ImageJ with "Show  
> Info..." (microscope settings, etc.) is also lost. The OMERO ImageJ  
> plugin does worse, as it loses the calibration altogether. The LSM  
> files themselves import fine into OMERO with all their metadata. But  
> we need to do some processing in ImageJ and save as TIFF before we  
> import the data into OMERO.
> - the image viewer within OMERO Insight is so slow it is unusable  
> when browsing slices of large stacks (~500MB). This is with the  
> client and server on the same machine (8-core MacPro with 16GB of  
> RAM); the machine was otherwise idle when I tested this. ImageJ is  
> much faster on the same stack, and I was wondering if it might be  
> possible to add a button to get a file to open in ImageJ (I know  
> this can be done indirectly by exporting the data file and opening  
> it from outside the OMERO client, but that's not very convenient),  
> or perhaps have ImageJ run within OMERO. Having ImageJ run within  
> OMERO would be very appealing because all the plugins out there  
> could be easily reused. Don't know if this can easily be done.
> - on OS X, the OMERO client has problems with file selection  
> dialogs. The list jumps to crazy places when trying to click on a  
> file name. The dialog also appears to be missing the standard OS X  
> shortcuts, which makes navigation very slow. Is there any way to use  
> native OS X dialogs? I imagine that's not a Java issue because it  
> appears that ImageJ does use native dialogs.
>
> Some suggestions for improvements:
> -  have relationships between images. For example, I'd like to store  
> one image and different segmentations of that image separately, but  
> not lose the information about the relationship between them.
> - have a "measurements" pane that appears along with image tags and  
> attachments. "OmeroTables" might be something somewhat similar, but  
> I haven't been able to try it yet (can't get PyTables working  
> properly on my machine). I need to define my own types of  
> measurements, which will be performed outside of OMERO but that I  
> need easy access to within OMERO.
> - the interface to get to the protocol/experiment editor is not very  
> intuitive. Also, would it be possible to get the protocol parameters  
> to display along with the tags and attachments, so one can quickly  
> see what they are?
> - cache image files on the client (I guess there are synchronization  
> issues, but image contents are very unlikely to change so that could  
> just be ignored as a first step). I see you have that listed in your  
> roadmap for 4.2.
> - have the client save the login password (or at least give the  
> option). It doesn't matter if it's stored in an unencrypted form. It  
> seems that forcing users to type a password everytime leads them to  
> either use a trivial password, or reuse the same password for many  
> different logins, both of which can be higher security risks that  
> storing passwords locally in plain text.
> - have a drop-in installation of the server, containing all the  
> necessary libraries. I imagine this might greatly extend your user  
> base, because you would then reach out to labs that don't have  
> personnel with system administration experience.
>
> Finally, the main concern one always has when spending lots of time  
> entering data with a piece of software is its long-term viability.  
> The data are stored in a completely open format of course, but still  
> converting them for a different piece of software might be tedious.  
> Do you have plans to keep developing OMERO in the long term, and are  
> you expecting the funding to be sustained?
>
> I'd like to thank you for your work, and for making it available as  
> free software. OMERO looks extremely promising, and I am sure it  
> will help set a new standard for experimental data organization,  
> archiving, and sharing between labs.
>
> Best wishes,
> Olivier Cinquin
> PS: all my comments apply to OMERO4.1.1 beta, running on OS X Server  
> 10.6.2

> From: Olivier Cinquin <ocinquin at uci.edu>
> Date: 5 January 2010 18:23:59 GMT
> To: Aleksandra Tarkowska <aleksandrat at lifesci.dundee.ac.uk>
> Subject: Re: Problems and suggestions about OMERO
>
> A suggestion I forgot to add is to have a "blind" view of a set of  
> images, i.e. have OMERO show a set of images without showing the  
> date and time of creation, the original file name, or the detail of  
> what experimental protocol they correspond to (for example whether  
> they correspond to treated or control samples). It might be  
> necessary to randomize the order in which the images are listed.  
> Users could score the features of interest blindly, and only see the  
> results once they're through scoring all images. A related  
> suggestion is to store all measurements on a per-user basis (i.e.  
> allow more than one user to define the same measurement for an  
> image), which would make it easy to see how well different users  
> agree in their [blind] scoring.
>
> Best wishes,
> Olivier



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openmicroscopy.org.uk/pipermail/ome-users/attachments/20100107/1592de47/attachment.html>