parameter

The -new option is the default for the parameter command if no option is provided on the command line. It creates a new parameter definition.

The primary purpose of parameter definitions are assign names to specific indices of the event's parameter array like object. This object is filled in by SpecTcl's application specific code in the event processing pipeline.

The names assigned by parameter can then be used later to define spectra, gates and other SpecTcl parameter dependent objects. Originally, SpecTcl parameters were only integer values and were designed to be values from digitizers that had some fixed number of bits. Over the years, and with the introduction by Daniel Bazin of the tree parameter package, it became clear that this was an overly restrictive definition.

Now, parameters in SpecTcl are double precision floating point variables. In systems with IEEE floating, this provides a sufficient number of bits of mantissa that integer parameters can still be faithfully represented.

Unfortunately, this step-by-step evolution, coupled with the strong desire to maintain backwards compatibility with user script has led to a rather confusing set of options for parameter definition.

All parameter definitions must provide a parameter name and a id. Note that using the tree parameter package allows SpecTcl to assign the id.

The simplest form of the parameter command is:

parameter ?-new? name id resolution

This creates a parameter that is intended to be an integer with resolution bits of resolution. Thus the range of the parameter is assumed to be from 0 to 2**resolution-1 inclusive.

parameter ?-new? name id resolution {low high units}

As an intermediate between integer and floating parameters, SpecTcl supported, and still supports, scaled integer paramers. These are integer parameters that map linearly into some floating point range.

The form of the parameter command above, defines a parameter with range defined as in the first form of the command, but whose values are assumed to linearly map to some real coordinate line in the range [low,high]. The command parameters low and high are floating point values. The parameter is also assumed to have units (e.g. cm or MeV) described by the string unit. Parameter of this sort are called mapped parameters and are useful when building histograms of simple calibrated parameters.

parameter ?-new? name id ?units?

Defines a real parameter has an arbitrary range, and optional units. The units parameter is a string that represents the units of the parameter (e.g. "MeV")

Returns a list of the parameter definitions. If the optional -byid switch is included, the parameters are sorted by Id number. Otherwise the list is alphabetical by parameter name. The output is in a form which makes processing by other Tcl commands and scripts easy.

The parameters listed can be filtered. If -id is not present, the optional pattern parameter is a glob pattern. Only names matching this pattern will be listed. If the pattern is not supplied, it defaults to * which matches all parameter names.

If the -id option is present, the next parameter is mandatory and must be the paramter id of the parameter you want listed.

The output is a Tcl List of definitions. Each element of the list is itself a list that describes one parameter. In order to make the processing of this list simple, all parameter definitions have the following elements (in order):

Deletes a parameter definition. Note that when objects that depend on a parameter are created (e.g. spectra), a binding is statically done to the slot in the rEvent the parameter is assigned to. This means these objects will continue to function. What they do when listed depends on the type of object.

If the -id is not present on the command, the name specifies the name of the parameter to delete. If -id is supplied, then id is the id of the parameter whose definition is deleted.