Version 2.8 of LaBunaPera writes data files in ascii or binary format. Details of these two formats are as follow. For both formats, the structure the same and includes a header and a footer which are always in ASCII. The two formats differ only for the representation of the Datalines, which, for long experiments, constitute the largest part of the file. The file is organized as: description (free text, even multiline) Date and time =============== Preset: presetVar1=value1; presetVar2=value2; ...; ========== Time; setVariable1; setVariable2; ...; getVariable1; getVariable2; ...; Dataline1 Dataline2 .... DatalineN ========== Post-set: postSetVar1=value1; postSetVar2=value2; ...; In the ASCII format, Dataline is an ascii string containing the numerical values of the time, of the set variables and of the get variables, each terminated by a semicolon (;). Get variables containing vector data are listed with all the measured values separated by one space. The time is printed in %.3f format and all other values in %g format. Each Dataline is terminated by a \n. In the binary format, Dataline has the following structure: \00 t S set1 set2 ... G1 get1 G2 get2 ... where: \00 is the null char; t is the value of the time, in float32 (real*4) representation S... is the number of set variables (collectively stored as a vector), as uint32 set1,2,... is the set value, as a float32, G1,2,... is the number of elements of the variable get1,2,... (>=1), as uint32 get1,2,... are the get value(s), as floats Dataline is not terminated, and the \00 heading the next line (or "="=61 for the last Dataline) act as a record separator. Note that LabView (at least on linux) writes big-endian data. Both these formats in principle support Datalines containing all the same number of variables, but eventually different number of vector get values on each line. Such a situation is uncommon, and therefore silently disregarded by my matlab reading routines, but is taken into account for future use.