jprenaud.info

At the moment, I am spending a lot of time finishing programming projects before I swap jobs at the end of September.

It’s never too late to learn something and I am discover graphical user interface GUI programming, something that always fascinated me but that I had no experience of whatsoever (apart from the work on the MS Windows version of CHASM which is very limited in terms of user interface).

I have started to develop a small GUI for the C++ slope stability program (called hydrostab) for which I got NERC funding. I chose to implement the GUI with the Qt toolkit which is cross platform (so technically it can be compiled for MS Windows and MAC OS X although I have only developed on Linux so far). At the moment, it looks like this:

It’s not much yet at all, it’s very busy, badly organised and please forgive me the fact that nothing is properly aligned etc… So far, I have tried to make functions work and been mostly familiarising myself with the toolkit.

The important thing is that the application is multi-threaded. I have implemented the time loop in a separate thread. It means that the simulation can be controlled from the GUI during runtime. So far, the simulation can be started, paused, resumed and stopped. The example above shows a simulation that was started, paused, resumed and the confirmation dialogue before stopping it for good. It works very well so far and it will be really useful in the future. Imagine, you could for instance:

1. start the simulation, pause it, look at the first computed results and amend the parameters if necessary. No more need to wait for everything to finish.
2. set up several simulations (several slopes or several configurations of the same slope) and run them all together in separate threads. This is something some of my colleagues interested in optimisation and econometrics might be interested.

In a way, it is related to some of the high performance computing work that I have been involved with for another project but the whole graphical aspect is very new for me.

There is still a lot of work to do but it going in the right direction I think. The next step will be to create a GUI for designing the slope so that the program does not rely on the CHASM4 data format which is very limiting.