classpath
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re[4]: Using OSG for Classpath


From: Peter Kriens
Subject: Re[4]: Using OSG for Classpath
Date: Tue, 1 Nov 2005 10:53:22 +0100

It may be easy to describe the modularization, it is much harder to
get a large number of companies to agree on this description. OSGi is
quickly becoming the defacto standard for this type of modularization.

The OSGi has so called execution environments which are machine
readable descriptions so I could use that to create the subsets for
these execution environments. I guess I can do this from the
monolithic JAR file without intruding on the normal development.

Lets see what comes out of that exercise.

Kind regards,

     Peter Kriens


AH> Peter Kriens writes:
 >> TT> I've seen the OSGi idea come up before, in particular on the Harmony
 >> TT> list.  I still don't understand what concrete benefit it provides.
 >> TT> Could you describe that?

 >> The OSGi gives you modularization. Instead of one big chunk, you
 >> get many smaller chunks with well defined dependencies. This is
 >> usually considered a good thing. Not only for your customers, who
 >> can use just what they need, but also easier development. You would
 >> also allow others to develop part of the solution. For example,
 >> Eclipse uses this model to allow thousands of developers to play
 >> together without having integration hotspots. Additionally, the
 >> OSGi provides "private" packages, which is an extra level of
 >> protection and helps future change. And last but not least, small
 >> is beautiful!

AH> But all that OSGi seems to provide is a way to _describe_ the
AH> modularization.  But that was never the problem: it's easy to do.
AH> There are lots of ways to do it.

AH> The difficult job is to _do_ the modularizing.  To decide which part
AH> goes where, and what should depend on what.  And to make the cuts that
AH> separate the dependencies.  It seems to me that OSGi modularization
AH> doesn't solve the real problem.  In fact, it doesn't even address it.

AH> Andrew.


-- 
Peter Kriens                              Tel +33870447986
9C, Avenue St. Drézéry                    AOL,Yahoo: pkriens
34160 Beaulieu, France                    ICQ 255570717
Skype pkriens                             Fax +1 8153772599





reply via email to

[Prev in Thread] Current Thread [Next in Thread]