I just realized that the last time I have blogged about my pet project, Gpredict, was back in February 2007. Back then, the big news was the Sky at a Glance function and there has in fact been a lot of evolution since then despite the big silence in the Gpredict Developer Blog. Let me try to sum up the progress from last two years as well as what you can expect to happen in the weeks to come.
First and foremost, I'd like to express my happiness with the fact that Gpredict is now included in most Linux and BSD distributions that I know of and is even available for Mac OS X via the Huntsville Macintosh Users Group – just how cool is that?
The development of Gpredict has been quite slow in 2008. I guess the most significant event was the release of version 1.0 beta 1 in September 2008, which included the first prototypes for radio Doppler tuning and antenna rotator control. These were using the new network interfaces of Hamlib, namely rigctld and rotctld. Both of these network daemons were in an early stage, so a high-level user in the form of Gpredict GUI was very welcome.
As it turned out, releasing a beta containing all this new code was a very good idea because the bugs and shortcoming of the hardware interfaces were very quickly discovered by people who could test it using real hardware. I could only test the radio controller using my IC-765 and FT-817 and certainly nothing more than the Dummy backend for the antenna rotator controller. while the antenna rotator controller was working surprisingly well, the Doppler tuning algorithm was way too primitive. It was designed to control the downlink only and it ignored completely any attempts to use the same radio device for uplink as well. Technically, this meant that whenever somebody was trying to transmit – be that simplex or full duplex – Gpredict got very confused trying to correct the downlink frequency. Obviously, the Doppler controller algorithm had to be improved to be a little more sophisticated.
The screen shots above show the very first prototypes for the radio and antenna rotator controllers.
After releasing the 1.0 beta 1 in September 2008, I got quite many positive responses so I continued to work on the code. Unfortunately, Oct-Dec was a very busy period and I did not spend as much time on Gpredict as it deserved. But, the good news is that in February 2009 I have resumed active work on it and many code changes have been committed since then.
Needless to say, much of the motivation came from the community and the high interest the beta 1 released has enjoyed. The beta release has been downloaded more than 1000 times and there has been continuous feedback in the form of bug reports, feature requests and lately even patches! You can see all the open bugs and feature requests on the new Gpredict project page on SourceForge.
The short term plan is to finish an improved version of the Doppler controller and release a new beta. The specification of the new controller algorithm is available in my personal wiki and the implementation is coming along as well. A few days ago I finished an updated UI that takes both the uplink and the downlink into account:
In the mean time, Stephane F8CFE was also fooling around with the Gpredict code and he even created a version that implements dual Rig (one for TX and one for RX) as well as a first cut at the transponder database:
So, I still have plenty to do and the next few days are dedicated to merging Stephane's mods with my own code and ideas and try to have a beta 2 release ready as soon as possible. Stay tuned for updates in the coming days and weeks. Gpredict 1.0 final is now closer than you can imagine ;-)