We have started a project of that kind for French: translation of the Qt 4.4 tutorial, some Qt Quarterlies, some Qt Labs, then a bunch of doc. We are using SVN for the first projects; for the doc, we used a totally different system: a DokuWiki, with a tool for importing pages from the doc (written by ourselves). We used a wiki rendering plugin to create the pages we serve to our visitors (basically, a PHP script calling the rendering engine and writing the result in a file).
Why DokuWiki? It does not use any database, so is easily hackable in every chunk of code. Also, it has a plugin system that makes writing new rendering engines really easy (for the other kinds of plugins, we're doing it ... well, not so beautiful a way).
Where you will have problems is pages you don't have translated. What we are doing is downloading it from Nokia's servers, then do some hack to have it render well in our template, and eventually cache it - it costs several seconds for pages like QWidget!
Also, there were things we did not like in the Qt doc: for example, you cannot ask for the page
Thank you for the pointer for Qt 4. If we move forward I can look at that.
I found an old entry in a www.qtcentre.org forum that says Qt 3 QTable does not support spanning columns which is what I ultimately need to do.
There is a third party widget QicsTable that should do the trick and I think it is still available for Qt 3.
Chris, I'm also very interested about this topic. I haven't done much DDD myself but I have read about it and done some toying.
One thing I toyed with was to create a Repository to fetch entities (json-objects from the internet). I found it hard to use QNetwork because the asynchronous nature of the module. Repositories are often modeled to return an entity and the entity can then be used to traverse the children etc.
If you're interested in shaping the future of Qt, I suggest you get involved there.
Essentially, Qt is moving towards being an open driven project, and as a result, the open infrastructure I talked about is going to have to happen.
The work I did on testing infrastructure for merge requests is also being worked on by some people at Nokia, so I've largely decided to not reinvent the wheel, and instead, to fit in with their efforts.