Based on some other posts going around, I think it's likely that HiveMind can become a Jakarta project, once the IP is fixed. The basic sentiment is that mandating that it move over to Avalon is effectively choosing sides, something Jakarta doesn't do. I'm still analyzing my own feelings about this ... intellectually, it may make sense to move the code into Avalon, assuming it will fit. But at another level, I'm very nervous about dilluting the code, and the focus, of HiveMind into Avalon, given that there are many fundamental differences in coding and design style. For me, joining Avalon and introducing HiveMind features into it represents a much higher level of effort. I'd prefer to present HiveMind as-is, and let the Avalon guys pull out what makes sense for Avalon (and the Spring guys, and the Pico guys).
It's kind of like fantasy sports cars. HiveMind is my own personal Ferrari. Avalon is, perhaps, a Diablo. At a high enough level, a Ferrari is interchangeble with a Diablo: four wheels, lots of power, no trunk space. Yet the subtle cues provided by a Ferrari (the sound of that engine, the shape of the hood, the color of the paint) excite me in a way that a Diablo does not ... and a Diablo fan may not be engaged by a Ferrari.
As I've said elsewhere, HiveMind represents my personal "hot points", the issues that are most important to me, reflecting my personal approach to coding and design style. Most of what's in HiveMind exists, in a different mix, inside JMX or Avalon or Keel, but in a different mix, that doesn't give off those same subtle cues that mean so much to me. For others developers, Avalon or Keel or Pico may provide the right set of cues, may be the most natural approach to getting work done. For myself, or those who have very similar sets of hot button issues, the others are just not quite enough, and HiveMind is the way to go.
No comments:
Post a Comment