Generalizing interfaces and interaction with QGraphicsItems across form-factor boundaries

[Marco Martin]

On Sunday 22 August 2010, Stefan Majewsky wrote:
> On Tuesday 17 August 2010 13:21:15 Marco Martin wrote:
> > hmm, this looks a bit like qml to me.
> > and yes i'm aware and agree on the concerns over its architecture, but
> > perhaps it should not be discarded before evaluating if it could have
> > some applications there
> 
> I am in the process of evaluating QML. I have only played with it a bit,
> yet I am sure that we could put it to good use.
> 
> Could you please expand on "concerns over its architecture"?

to me is mainly the implementation of a new base class, called 
QDeclarativeItem that is almost-but-not-quite QGraphicsWidget.
(and also the layout system is different, while is great is some cases, it is 
less in others) to me makes it quite challenging for doing rich applications 
ui.

anyways probably this affects games in way less extents, so it could be good 
enough (or you may find completely different issues that are strictly game 
related, i don't know :p)

> > paint rutines: we have a quite efficient svg loading and painting
> > mechanism with on disk cache of the rasterized stuff that seems to scale
> > pretty well on not-so-capable devies (when there aren't cache misses it
> > is possible to display svgs without instantiating any svg parser)
> 
> As I mentioned in my reply to Aaron, incidentally I implemented a similar
> solution for kdegames just at the beginning of the 4.6 cycle.

yes, looks quite similar indeed, looks like we sould had this thread way 
before eheh :)

Cheers,
Marco Martin