Chapter 2. Overview of how SpecTcl works
The program name, SpecTcl, says a lot about how SpecTcl is built. Spec, stands for Spectra and Tcl for the Tcl/Tk scripting language. SpecTcl implements an extension to that language as its command language. This allows users to not only type simple commands at the program, but to write scripts of commands and, using the Tk toolkit, easily write graphical user interfaces that bolt on top of SpecTcl.
SpecTcl is often compared to Root, although our application domain is intended to be quite different. SpecTcl concentrates on making it easy to create histograms, define condtions (gates) and apply those conditions to spectra. That's its focus. Root is a general purpose anaylis framework of classes.
By confining our problem domain to making histograms, we can simplify the process of tailoring SpecTcl to a specific analysis task. In Root, for example, you are typically required to write the entire analysis from raw data to desired output products. Since SpecTcl knows that our output products are just histograms, we can relieve of the task of creating and incrementing histograms.
Here's a simplified picture of the flow of data through SpecTcl:
Events packaged in accordance with the format of some dat acquisition system's outer format arrive at a SpecTcl Buffer Decoder object. The buffer decoder is responsible for knowing about the outer format of the data (sometimes called the data envelope). It opens those envelopes and produces events.
A logical pipeline of event processors is given each event. The event processors in this pipeline are the only bits of software needed to tailor SpecTcl to a specific analysis problem. Each stage of the pipeline has access to the raw event and the parameters that have been produced by previous stages of the pipeline.
The parameters that have been produced by this event processing pipeline are then passed on to the histogrammer. The histogrammer uses a set dictionaries of parameters, spectra gates and gate applications to, given the parameters of each event, determine how to increment the spectra.
A visualization component of SpecTcl provides a view into the histograms, gates and their applications to the user.
Several important points about SpecTcl:
Event processing is done as background processing from so the user interface, including the command interpreter, is live during event processing.
SpecTcl can accept event data from two types of sources; files and UNIX pipes. The unix pipe data source allows SpecTcl to be attached to online data acquisition systems to do analysis of live data.
Creation of histograms, gates and application of gates to spectra is totally dynamic and can occure while analysis is in progress.