CowTalk

Here is another new interesting new hopefully useful project for developers of distributed Java systems: Dirmi. It is essentially a replacement and upgrade for plain old Java RMI, and both addresses most existing issues with vanilla RMI and extends set of functionality. And does all of that with a few usability improvements. Sounds pretty good to me.

So why do I think this might be a very good replacement? Beyond reading its feature set, I have not yet really used it. But I have confidence knowing its author, who has written such solid packages as Carbonado (ORM to use with BDB, amongst other backends) as Cojen (code generator). I hope Dirmi will get bit more exposure in near future (as well as Carbonado that seems to be somewhat of a well-kept secret and deserves to be more widely known). I will try to write a follow-up if and when I get to play with Dirmi a bit.

Posted by Tatu Saloranta at Wednesday, November 11, 2009 10:36 PM
Categories: Java, Performance
| Permalink |Comments | links to this post

Ok, I had seen "SOA Manifesto" link on my Google reader (written by Stefan T), but hadn't paid much attention to it. Too busy designing and implementing services, as it happens. But today I decided to have a peek, and was pleasantly surprised. I don't know what's with manifesto's, but it seems like ones I bother to read make sense. Granted: my sampling of manifestos is little more than sequence of the venerable Clue Train Manifesto, and Agile Manifesto. But the track record here is bit better than with other kinds of declarations within technical domain.

At any rate, go ahead and read it, it belongs to Just-Makes-Sense category. Good stuff, pragmatic, not too fluffy. Sometimes it does really pay off to find bit of common ground, in between dead-horse-beat-fests.

ps. for an alternative view, you can find many rebuttals (like this one -- but don't waste too much time on it; it's not written very well, and mostly focuses on strawman argument ["it can't be good because it was written by bad guys wrote crappy software"]).
This discussion (*) seems to be one of those cases where best arguments for something are made by people who are against the thing. :-)

(*) Assuming you can call talking past each other a discussion

Posted by Tatu Saloranta at Monday, November 09, 2009 11:39 PM
Categories: Philosophic
| Permalink |Comments | links to this post

Looks like some brave folks are using Jackson on Google's App Engine (for JSON processing). Neat. Some minor problems exist (need to add a patch similar to one suggested), and chances are this is not the only thing one can encounter. But there shouldn't be anything fundamental difficult, based on succesful usage.

Oh, and on related note: Jackson 1.3 was just released. Check out feature list for details; list is long -- lots of fixes, incremental improvements -- but this time around there weren't many big features, just plenty of bug mop-up, added convenience methods, extended support for partially supported things and so on.

UPDATE, 06-Dec-2009: There have been some important fixes to issues runing on GAE and Android platforms -- make sure you use version 1.3.2 instead of 1.3.0 if running on either one (and please add comments if you do; whether it works, or you find other issues: sooner problems are reported, sooner they get fixed)

Posted by Tatu Saloranta at Wednesday, November 04, 2009 11:57 PM
Categories: Java, JSON
| Permalink |Comments | links to this post