[ome-devel] [ome-users] RFC: Changes to the OMERO minimum system requirements for 5.1 and 5.2

[Jason Swedlow (Staff)]

Hi Lee & Roger (and everyone else!)

Thanks to Roger for putting this info all together and driving the discussion forward.  Anyone who is watching this and our sister open source software projects knows these issues are looming large in several ways.

Many thanks to Lee for the detailed feedback- is very helpful.  We've received a few private responses (we prefer a public discussion, but for some entities these things have to stay private) and obviously there are discussions on other lists, which we are keeping an eye on.

In general, any and all feedback is most welcome.  It does seem that many of us are dealing with the vagaries of versioning and compatibilities of Java and OS X, and Python, and bitness, and....

Thanks again.

Cheers,

Jason


From: Lee Kamentsky <leek at broadinstitute.org<mailto:leek at broadinstitute.org>>
Date: Tuesday, 13 January 2015 14:27
To: Roger Leigh <r.leigh at dundee.ac.uk<mailto:r.leigh at dundee.ac.uk>>
Cc: OME Development <ome-devel at lists.openmicroscopy.org.uk<mailto:ome-devel at lists.openmicroscopy.org.uk>>, OME Users <ome-users at lists.openmicroscopy.org.uk<mailto:ome-users at lists.openmicroscopy.org.uk>>
Subject: Re: [ome-users] [ome-devel] RFC: Changes to the OMERO minimum system requirements for 5.1 and 5.2

Thanks, Roger, I hadn't planned on replying to this, but having a detailed proposal brings up some questions that are both relevant for our OME support and more broadly for other community tools (like CellProfiler).

* Centos versions: I'd like to think that my organization is at the forefront when tracking technology, but we are currently on Centos 6 for our cluster. I think we'll be moving to 7 within the next year, but I can see other institutions lagging behind. One positive is that we are moving toward cloud apps and the applification of OMERO is a huge win that could make the version problem a moot point.

* Bitness - On one hand, I would love to discard all our 32-bit builds. On the other, moving the Mac build of CellProfiler to 64 bits has sadly been an overwhelmingly challenging task due to migration from Carbon to Cocoa and the difficulty of integrating a Python and Java UI on that platform. Although solving this is a priority, dropping 32-bit client support on the Mac could leave us behind on that platform. But I understand the challenge of maintaining a 32-bit build and realize that it's up to us to do what's needed to keep pace.

* Python - I can see no compelling reason for our tools to move to 3.x and we will probably stick with 2.7 or 2.8 if it comes out as long as the packages we need support 2.x. We have dropped 2.6 support.

* MSVC - you can't buy MSVC 2010 professional any more, although that is the compiler that's best supported for compiling 64 bit extensions for Python 3.x. We actually use the Microsoft SDK to compile CellProfiler - that uses MSVCR90.dll which is one step behind MSVC 2010, but is the implementation of the C runtime library used in Python 2.7. One important task for us is to learn how to build Python extensions using MSVC 2013 to produce assemblies that use MSVCR90.dll instead of the default for MSVC 2013. But, I think it's important that you use tools that people can actually get, so I would be in favor of supporting the latest and I can understand dropping support for earlier as long as any binaries loaded by Python are truly compatible. Informally, I've patched Python's distutils to make it capable of using MSVC 2013, but I think it needs further patching to solve the DLL problem.

* Java - we currently build on 1.6 but we would like to track OME and ImageJ and we would welcome a community decision to build on 7. The situation on OS/X 10.7 and Java upgrades on OS/X in general is a little daunting for some of our users, but I think the community consensus will tell us (CellProfiler) what's best to do, so we appreciate the insight and direction that you're supplying.

--Lee

On Tue, Jan 13, 2015 at 8:46 AM, Roger Leigh <rleigh at dundee.ac.uk<mailto:rleigh at dundee.ac.uk>> wrote:
Dear all,

Prior to the release of OMERO 5.1, the OME team have been reviewing our
support for the various software components we depend upon.  A proposal
of the versions we intend to support for OMERO 5.1 and 5.2 is detailed here:
https://www.openmicroscopy.org/site/support/omero5.1/sysadmins/version-requirements.html

Our resources to test and support multiple versions of each component
have limits, and older versions of components lack security support and
updates.  We need to deprecate and drop support for older versions over
time, as well as limit the number of versions we support at any one
time.  Versions for which we will drop support for in 5.2 are marked as
deprecated in 5.1.  Notable here are the deprecation and dropping of
support for Java 6 and older PostgreSQL versions, but please look at the
above link for full details.

Your feedback would be appreciated.  If any of these proposed changes
are problematic, please do provide us with the details of your system
(OS version and software versions) so that we can try to accommodate
support for it where possible.


Regards,
Roger Leigh

--
Dr Roger Leigh -- Open Microscopy Environment
Wellcome Trust Centre for Gene Regulation and Expression,
College of Life Sciences, University of Dundee, Dow Street,
Dundee DD1 5EH Scotland UK   Tel: (01382) 386364

The University of Dundee is a registered Scottish Charity, No: SC015096
_______________________________________________
ome-devel mailing list
ome-devel at lists.openmicroscopy.org.uk<mailto:ome-devel at lists.openmicroscopy.org.uk>
http://lists.openmicroscopy.org.uk/mailman/listinfo/ome-devel


The University of Dundee is a registered Scottish Charity, No: SC015096
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openmicroscopy.org.uk/pipermail/ome-devel/attachments/20150115/aca73dda/attachment.html>