The downloads are behind, as the status says: for the moment, be prepared to get it via anon CVS and build it using Maven.
The community is re-forming right now; a vote for an alpha-4 or a beta-1 will be forthcoming. -- Howard M. Lewis Ship Independent J2EE / Open-Source Java Consultant Creator, Tapestry: Java Web Components Creator, HiveMind http://howardlewisship.com > -----Original Message----- > From: Dieter Bogdoll [mailto:[EMAIL PROTECTED] > Sent: Thursday, April 15, 2004 11:15 AM > To: [EMAIL PROTECTED] > Subject: Hivemind > > > Hi Howard, > > I wanted to have a look at Hivemind, but according to the > download links > the latest version of HiveMind is 1.0 alpha 3 from 23.Oct.2003. > Is the project currently alive, or is it hibernating? > > I have some questions: > > 1. regarding the following hivemodule.xml fragment > > <service-point id="Add" interface="net.bogdoll.hivemind.IAdder"> > <create-instance class="net.bogdoll.hivemind.AddImpl" > model="pooled"/> > </service-point> > > Although is specified the model to be pooled no activateService or > passivateService method will be called. > According to the docs on your web site thats the right syntax (and it > will be parsed correctly). > After browsing your code, I found, that it would work if I > put the model > attribute into the service-point tag. > But I think the model should be with the implementation and > not with the > role or interface. Since alpha-3, the model attribute has moved from <service-point> down a level, to reflect that the model is related to the implementation, not the interface. > > 2. I liked the lookup and release style of the avalon > service/component > manager. With hivemind I have to fetch the ThreadEventNotifier and > call there the fireThreadCleanup method. Okay that works, but > it is not > what I semantically want. I just want to release the hold > onto a service, > this has nothing with Threads to do. I'm picturing very large systems with dozens if not hundreds of services. It would be very tedious to have every point of access into the system have to know which services used which service models and needed to be released. In HiveMind, the proxies that you get back from the Registry are fully threadsafe; they can even be shared between threads and still operate correctly (each thread has access to just that thread's data). However, there's no magic incantation I know of to determine when the current request ends. For web applications, we have a servlet filter. I think in the future we'll see other, similar adaptors for things like AXIS and other likely points of entry into HiveMind. However, this does beg the question: perhaps the pure division by *thread* is not appropriate. Maybe we should have a "transaction", "session" or "unit-of-work" metaphor instead. That's why HiveMind isn't a 1.0 release yet ... we're still deciding on which approaches yield the best results. > > 3. What happens if I put two implementations to the same > service-point? > First wins? Yes, but with an error for the second implementation. > > I like your framework and I hope it is still alive. HiveMind is gonna rock and rule for a while yet! > > Bye > Dieter > > -- Howard M. Lewis Ship Independent J2EE / Open-Source Java Consultant Creator, Tapestry: Java Web Components Creator, HiveMind http://howardlewisship.com --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
