Archive for November 28th, 2005

Posted on November 28, 2005 - by jono

The future is Cairo

This week I am giving a PHP and MySQL course, hence being rather busy teaching my 8 young padawans the finer concepts of web development. We made some good progress today. It is always fascinating doing these courses as they give an essential, grounded insight into the variety of reasons why people want to learn the technology. I also love to see those eyes light up at what they can do…

There has been some interesting development recently with regard to Cairo. As many of you will be aware, the idea of Cairo excites me to unnatural proportions. Aside from the obvious graphical improvements that are inherent in a finer grained and better quality rendering system, I am really excited by the way Cairo opens up more possibilities for artists and designers. We really need more of these people on board – we are seriously starved by the lack of quality artists available, with only a few at the last headcount. With all the discussion of how distributions such as Dapper should be polished and sheened, a lot this effort needs to go into the visual quality of the interface. We really need teams of dedicated developers who are fluent in Cairo to turn our applications into truly beautiful interfaces.

A good example of this is MythTV. Although I am pant-wettingly excited about my snazzy MythTV box, the interface really could do with some work. The main problem is that the interface is not particularly optimised, and although it looks good, it could look so much better and work so much better. This will take time as it is written in Qt3, and a Qt4 port is not on the cards until Isaac feels like the Qt has stabilised a few point releases down the line. The problem of course is that Qt does not naively support Cairo, and although there are examples, it is going to be more of a pain to use Cairo extensively without native support. This naturally requires some buy-in from Trolltech, and I really do hope they go for it. Can anyone shed any light on this?

Recently though, it seems more and more applications are moving to Cairo, such as GTK, GNOME themes, Diva, OpenOffice.org etc. I don’t think many people deny Cairo the crown as the vector graphics king, and despite some concerns about speed issues, it seems the developers are really ploughing through the bug reports and fixing up performance as best they can. I would be fascinated to see someone like Federico or Behdad go to town on Cairo and see how it can be optimised, but then again, I don’t know if you need to know about low-level maths algorithm hacking to optimise it any further.

I really want to have a crack at doing some Cairo sometime soon. The quality of work that some people are coming out with, such as Clearlooks is incredible.

The big question is, if Qt supported Cairo as the backend, would this make themes work for both Qt and GTK with minimal modifications?