Java 9, OSGi and the Future of Modularity

InfoQ today published the first part of a two part article written by Neil Bartlett from Paremus and Kai Hackbath from Bosch / ProSyst.


Its definitely worth a read. The article was written some time ago so is in advance of the latest delay to JPMS / Jigsaw (and consequently Java 9) that was proposed last week on the mail list and at JavaOne this week.  This is just the latest in a series of delays over many years and we suspect won’t be the last.

As a very quick synopsis the conclusions are:

  • Jigsaw / JPMS has taken a very long time and had lots of false dawns.
  • It has done a good job of modularizing the Java platform itself (we can construct small runtimes containing only relevant parts of the Java platform for a specific workload)
  • It has underestimated what is needed for application modularity. JPMS comes up very short when compared with what OSGi has been in doing successfully in this area for many years.
  • The combination of JPMS for the Platform modularization and OSGi for application modularization looks set to be a winning formula, just so long as the JSR doesn’t break OSGi in its dogged attempts to do application modularity ‘better’.

But don’t take my word for it.  Please have a read for yourself and lets us know what you think…

And of course Part 2 of the article will be coming up in a couple of weeks so be sure to watch out for that final instalment too.

Share This:

Leave a comment

Your email address will not be published. Required fields are marked *