Keep raw data and processed data separate

[cjlegg]

Only tool srcountsanalyzer has an outputdir option. This is not the only difficulty. The inputdir an outputdir options for all the tools should be reviewed. In particular it should be recognized that the raw data exists in at least two directories - the main one backtrace directory - in fact it is possible that the user could elect to have a separate raw data directory for each of the five call sampling libraries. Further it should also be taken into account that the tools produce intermediate data files from the raw data which are then used as inputs to another program(the .png graph in the doc directory shows the intermediate data files as the ones in the middle row). The options should allow the user to take in inputs from all these different directories and direct them to an output directory that they wish (or it should stated that the output directory is the command's working directory). Also it should be possible for the user to keep the intermediate and final data sets in separate directories from the raw captured data to avoid confusion and easy re -analysis.

[gvallee]

This is not a product, it does not matter to me.

[gvallee]

After discussion the point is more about keeping the raw data pristine and have the processed data in different directories

[gvallee]

A first step would be to have all the command line tools support the specification of the output directory.
Of course this means that some tools will need two inputs: the raw data directory and the processed data directory.
We also need to audit the various command line tools to know which tools already does what.