[ome-devel] Opinions on Java 5

[Ilya Goldberg]

To clarify, discussion of the relative virtues of different programming 
languages is academic and largely moot (and leads to flames and 
religious warfare).  The simple reason why this discussion is focused 
on Java is that there is a great horde of Java developers, and as a 
consequence a great variety of open-source projects for libraries, 
frameworks, development tools, etc.  The drive for the discussion was 
to try and farm out components of OME to other independent projects in 
order to reduce the complexity of our code-base (thereby reducing the 
cost of maintenance), improve performance, and "features".  
Specifically, we would like to farm out the object-relational mapping, 
and both ends of the middle-ware (remote objects) without imposing 
language requirements on the client, while preserving the dynamic 
nature of our data model.  I'm not at all familiar with the Ada 
community.  Are there open-source Ada projects that already do these 
sorts of things?

Take care
Ilya

On Aug 19, 2005, at 4:41 AM, Chris Allan wrote:

> Bob,
>
> This isn't an Ada discussion, it's a Java discussion. We have no
> current interest in Ada.
>
> Ciao.
>
> -Chris
>
> On Thu, Aug 18, 2005 at 09:02:50AM -0700, Robert C. Leif wrote:
>>    Although Java has some benefits over C and C++, its major virtue 
>> was a
>> brilliant sales campaign. Java was introduced without generics or 
>> enumerated
>> types.  I gather that generics, which are extremely useful, have 
>> finally
>> been added.  Does Java 5 have enumerated types?  Aonix
>> (http://www.aonix.com/) has an add-on for hard real-time that should 
>> be
>> useful.  C# is a similar effort by Microsoft and ECMA.  C# does not 
>> appear
>> to be used much, if at all, on other operating systems.
>>    Since some of you might not know that an alternative to both 
>> languages,
>> Ada, exists, I have added the following to provide an idea of what 
>> should be
>> expected in a programming language.  Ada has portability including 
>> maximum
>> operating system independence, ISO standardization, excellent 
>> generics,
>> safety, readability, superior real-time performance including 
>> rate-monotonic
>> scheduling, interfaces to C and other languages, and good error 
>> messages.
>> Obviously, there are advantages for a GNU compiler.  In fact the best
>> implementation of Ada is the GNU compiler, GNAT.  I do not know about 
>> other
>> GNU compilers.  However, the support for a free compiler can be 
>> greater than
>> the price of a commercial compiler with support.  For instance AdaCore
>> (http://www.gnat.com/) prefers to sell support for their Professional
>> (commercial edition) for $10,000 for four seats per year.  The free 
>> version,
>> 3.15, of the compiler is about 3 years old and the free version of the
>> environment is over one year old.  Academics can get the Professional
>> Edition free.
>>    Bob Leif
>>
>>    -----Original Message-----
>> From: Josh Moore [mailto:josh.moore at gmx.de]
>> Sent: Thursday, August 18, 2005 1:10 AM
>> To: ome-devel Development
>> Subject: [ome-devel] Opinions on Java 5
>>
>>    -----BEGIN PGP SIGNED MESSAGE-----
>>    Hash: SHA1
>>
>>    Seems like I haven't started a good heated discussion in a while so
>>    anyone got an opinion:
>>
>>    	Does it make sense to use Java 5 on the server side?
>>
>>    This ties in to the "Explorations of alternative architectures for
>>    remote clients" thread. The code I've written currently has 
>> dependency
>>    on Java 5 _for the server side_.
>>
>>    However, this makes server development on OS X 10.3 impossible and 
>> on
>>    10.4 apparently difficult. It also complicates installation, at 
>> least
>>    for most people at the moment. (Java 5 uptake has been relatively 
>> slow.)
>>
>>    The pros of using it are:	
>>     - generics
>>     - a few API methods and non-thread safe objects
>>       (StringBuilder rather than StringBuffer, etc.)
>>     - possible use of annotations (e.g. 
>> http://annotations.hibernate.org/)
>>     - ...
>>
>>    The performance gain doesn't really count because one can use the 
>> new
>>    JDK but still compile for JRE>=1.4 (or maybe even 1.3)
>>
>>    However (again) if we don't do it now, we'll need to do a "port" at
>>    sometime in the future.
>>
>>    Ho hum.
>>
>>    Any thoughts?
>>      Josh.
>>    -----BEGIN PGP SIGNATURE-----
>>    Version: GnuPG v1.2.5 (GNU/Linux)
>>    Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
>>
>>    iD8DBQFDBEJqIwpkR5bKmAsRAsmqAJ9Pk0rP+sowiy7GBegGmlC03hek3ACeLfmu
>>    ymIp+hW3/d2X7Oii2eKs3DM=
>>    =TRHx
>>    -----END PGP SIGNATURE-----
>>
>>
>> _______________________________________________
>> ome-devel mailing list
>> ome-devel at lists.openmicroscopy.org.uk
>> http://lists.openmicroscopy.org.uk/mailman/listinfo/ome-devel
> _______________________________________________
> ome-devel mailing list
> ome-devel at lists.openmicroscopy.org.uk
> http://lists.openmicroscopy.org.uk/mailman/listinfo/ome-devel
>