There is a GUI-based event display program for the Radphi experiment called Event. Event was developed at JLab by Troels Petersen under the supervison of Elton Smith, with help from Dave Heddle and James Muguira. The latter two are the authors of the underlying graphics package called Hv [1], an abbreviations for Hot views (its not what you think). Hv is a rather sophisticated package (i.e. complicated) with many features beyond what we really need for looking at Radphi events, but it worked rather well for the 2000 run and many of us got used to it. Maintaining and porting Hv to the latest version of Linux or solaris is beyond the resources of Radphi, but CLAS has now incorporated Hv into its suite of centrally maintained libraries. With the prospect of CLAS maintaining Hv, it makes sense for Radphi to maintain Event as one of our offline analysis tools. This web-based log book is devoted to that ongoing project.
Back during the summer when I was first porting the main components of the Radphi software repository to Redhat Linux, I ran into the problem that the Hv library we were using under Solaris was way out of date. After hunting around on the JLab disks I found a version of Hv for Redhat 6.1 in the CLAS repository. This was good news because it means that we can get the Hv binaries from CLAS and avoid having to do the maintenance ourselves. But there is bad news too. The version of Hv that Event works with is so old that it doesn't compile with the current Hv library. I don't remember the details from last summer, but either the argument lists for some functions changed, or some arguments that were not checked before are checked now or something, but the net result was that the root Hv window would not even open. The whole color scheme has also changed, which required considerable reworking of the code that converts ADC and TDC values to colors. That much I did get working last summer.
Then I discovered that a lot of short cuts had been used to make Hv work (in a big hurry) last summer. So things that we are now doing in makehits, like pedestal subtraction, gain balancing, TDC window cuts, etc were being done independently in Event, mostly with hardwired constants in the code, etc. Unfortunately it has proved quite time consuming to remove all of this stuff from Event and tell it instead to get its information from the hits arrays. That is what I am working on now. Once that is done, we will be able to use the event display to check the correct operation of the makehits routines, and to validate the Monte Carlo and higher-level analysis tools we are developing. So far I have not yet completed this step.
Wish I had known that a month ago, when I spent time trying to make Hv run without core dumping. Right now when I try to read Monte Carlo into Event, it starts up OK and then does a segmentation fault as soon as I request the first event. Any ideas what causes that?