On Hivemind itself I must admit I used it, I loved it, and I have moved on. It was (still is) a brilliant conceived IOC framework that pre-dated and outclassed most of its contemporaries (e.g. Spring). Its key killer features were
- Auto-wiring that works - it made auto wiring in general possible.
- Proxying objects for different lifecycles - the ability to trivially create per thread, per session (and per anything proxies) is immensely powerful
- Contributions - They took a while to get your head round but they are an ideal way of telling services how to work
Obviously now it has been superseded by the equally superior tapestry-ioc. The does all of the above and removes the XML configuration.
Regarding the idea of Apache Attic - I think this is an excellent idea. Apache has been suffering for several years with a number of deceased or dying projects and this gives them a way out. If I had to suggest some I would suggest they consider moving
- Apache Turbine - I used to love this - but it hasn't changed for years. Other things (e.g. Tapestry) have come along that do what it does and a lot more.
- Apache Jetspeed - A interesting portal - but hasn't released since 2007
- And I am sure there are a load more...
It would be nice if being an Apache project meant (like it once did) having a substantial community and active development!
Posts
1 comments:
HiveMind just lost momentum when I shifted focus to Tapestry 5 IoC. Originally, Tapestry 5 was to be built on HiveMind ... but as I was writing early documentation for T5 I went from a paragraph extolling the evil of using XML configuration and then shifted directly into a parameter that started "now open up hivemodule.xml and ..." ... and realized that HiveMind had to go as well.
I feel bad for the developers who had committed to HiveMind, and for James Carman, who I convinced to take over the PMC for HiveMind. I really feel like I "pulled a Strachan" on this one, but it could not be avoided.
Post a Comment