Xulux is a typical tool that will never see a release unless I am really happy with it.
If people are going to use xulux, they just have to be sure that I will be able to support it, else the choice of using xulux as the gui framework (or frontend API, as I like to call it), will be risky business.
With the current state of xulux I've hit the limit of what I can do and keep backward compatible with the code we use at the office. I've learnt a lot from development of xulux and experienced first hand the down side of the approach I took. Creating gui's is very fast in xulux (at least for the users of xulux), it saves a huge amount of time, but adding business logic and the interaction between the widgets is not that great. Almost everything can be done, but it takes more coding than I had in mind. Also extending the widgets is pretty copy & paste a complete widget, where you would actually like to just extend part of the functionality.
So I started on xulux2 last week. There will be no release of xulux "1" at all, just the snapshots (which works great for me). Parts of xulux1 will serve as a base for xulux2 (mostly the utils,swing code and the bean data provider). The frontend API will be similar but a lot smarter.
Hope to get a lot done the coming weeks (although I am currently in a restructure of my house, which takes a lot of my time too).
Will try to keep the core to at least support jdk 1.3 and my initial goals are supporting swing, swt and servlets, add scripting support and add support for (custom) rule engines like drools.
I seem to end up as the guy who actually doesn't release anything. I create a lot of (open source) stuff, but I cannot get any further than alpha releases (ehh SNAPSHOT releases). Never releasing anything is, viewed from my position, has a pretty clear reason : If you release you have to be prepared to actually support things (at least that's the way I feel about it). I don't mind giving support, but having to support your software, prevents you from doing actual development and making progress (although it may vary on the tool you are writing).
Since we are talking opensource you would expect that a good reason for releasing is that other people may get involved in the development effort and this way you can do more and get better products.
Yep it COULD. If it doesn't you have to support the stuff on your own, for a system that you are not yet ready to support.
I don't write open source specifically to make other people happy, just to make myself happy. I just try to solve problems I hit and if the result will help other people, I am happy for them.
The reason why I still choose the open source model for my projects depends on where I hit the problem. Open source is an excellent way to make sure that your investment in building the tool that will solve problems in your day job, will not be lost when you eg switch jobs (I tend to put a lot of "spare" time for my day job to do stuff like that).
In short : If I am going to release something , I am prepared to support it and I am very happy with the product.