Hi all, I'd like to ask to you a crazy idea I had in last weeks ... you remember the question of last week on add some methods to Pivot collections ? It's here: http://apache-pivot-developers.417237.n3.nabble.com/No-quot-addAll-quot-method-on-ArrayList-td3281354.html
After thinking on it, and trying to expand the idea (and you know that I'm a great fan of Scala) ... so what do you think (for the 2.1 release) to add some methods to pivot collections to be able to pass a Function (a Pivot new class to map simple functions, all for Java ... and the name is only a proposal) and have the collection to apply it over all their elements ? Do you think could be interesting/useful, at least for what it's your experience on Pivot ? If you have some time/are interested, try to take a look for example at these articles, just to have some idea: http://markusjais.com/a-little-functional-java-the-map-function/ http://www.ibm.com/developerworks/java/library/j-scala06278/index.html http://functionaljava.org/examples/1.5/ And note that Java 8 will have something like this (but less powerful to Scala, for what I have read ...), and JavaFX 2 will have something already aligned with it ... so I'm curious to see what will happen there :-) . But before it I'd like to hear from you what you think ... On Scala integration I have prototyped something that converts between Pivot and Scala collections using implicits, but I don't know it this is enough in a long-term vision (and even from a performance point of view, and on this I have to ask to some Scala expert), so the following idea (ok, crazy) but for Pivot 3: moder JVM language like Scala and Groovy have many types of collections that are event-enabled (and for example in Scala it's possible even to add methods to existing classes), so an idea could be to change Pivot collections to align (for what it's possible) with more robust/powerful collections (as method names etc), and retrofit them with interfaces that we will use even to extend/wrap other collections, and use factory classes to get instance of right collection, depending on some conditions (technology used, environment, etc). And maybe useful even a singleton class the Pivot classes could use to share settings etc in memory (for example factories, but maybe even other things) ... this could be useful even in 2.1. I know, it's a lot of work, but I think could be powerful ... what do you think ? You can say to me that I'm crazy, don't worry :-) ... Note that of course I'm not saying here that Pivot collections aren't good (without doubt Greg has made a huge/great work, you know I have maximum respect for it), but only that things are evolving, so maybe could be good even for us to start to align with new trends ... To finish, I dropped the pivot-scala project, and the new pivot-multilang is defined as structure but contains only few things because hadn't time to migrate existing scala stuff in it ... it's the first in my TODO list after the bug fixing stuff and the jenkins builds ... If you think it's an interesting idea I'll prototype ... Thank you. Bye, Sandro
