Running an Experiment

The standard way to start a PXL program is to type the program's name followed by an option for the parameter file to be used and the subject code:
   icon -f sperling.x john
will start the program icon and tell it to use johnas the subject code and use the parameter file sperling.x. The subject code is optional. If it is missing then the program will use vp as its default subject code. In this case the program's behavior concerning the creation of data files is somewhat different from the case where an explicit subject name is given. The subject code is used to create a file name for the data output files.

For any program there is at least one data output file. It contains the complete and detailed information on each trial and each experimental variable. This file gets extension dat. Depending on the program type there may show up another data file using the extension res. Creating this file usually is accomplished by defining the experimental variable resformat to output results in a certain format (see Chapter exvars for a description of resformat and for the mechanism to use it. A 3rd data file may exist for giving results of computations that are done during or after data collection. This file gets the extension pdt. The data file with extension dat is written while data are collected. The other results files are usually not written before data collection is finished.

Any default file creation can be overridden by giving an explicit command line option. Options follow the program name and are introduced by a minus sign (-). Here is the command line syntax

   command [options] [subject-code]
The command name depends on the program name, it was icon in the above example. The subject code is used to identify the subject as described in the previous sections. It was john in the above example. Here is a complete alphabetical list of all options which are known by any PXL experiment:

-C
write a parameter definition code file.

-c
do not run the experimental sessions, computations only. This option should be used if one wants to use a dat file as an input file to do some additional computations that have been suppressed at run time.

-D macrodef
send the macro definition macrodef to the preprocessor. This results in the expression

#if	defined(macro)
evaluating to TRUE. See Chapter macros for an example showing how this feature may be used to define parameter values on the command line.

-d n
set sounddevice to n. Available sound device numbers are listed in Chapter sound.

-E filename
use filename for error output.

-e
use stderr for log file output.

-f parfile
use parfile as the parameter file. This overrides any default.

-h n
set screenheight to n.

-I path
add path to the list of include file paths. This option may be given only once in the command line.

-i
present an instruction file to the subject at the start of the session.

-L n
set active language (n=0 is German, n=1 is English).

-l filename
use filename as log file.

-m
suppress macro preprocessing. This option may be necessary if a PXL program is to be run under a debugger. Since the macro preprocessor is run as a subtask it needs a considerable amount of memory. This will usually not be available if there is a debugger in the background. If there are macros in the parameter file one may use prepxl.exe first to preprocess the parameter file and then use the preprocessed file as a parameter file and give the option -m to save memory and time.

-o datfile
write data on file datfile.

-p pdtfile
write computation results on file pdtfile.

-Q n
set printflag to pattern n.

-q n
use debugging mode n.

-r resfile
write results defined by resformat on file resfile.

-S
simulate subject responses (not always available).

-s
use stdout for data and computation results output.

-t n
set switchtype to n. See Chapter timing for a list of available switchtype values.

-v n
set videomode to n. See Chapter graphics for a list of available video modes.

-w n
set screenwidth to n.

-X n
return n as exit value.

-x
do not run the computation phase, experiments only.


Back to table of contents


Author: Hans Irtel

irtel@psychologie.uni-mannheim.de