I had to discover how par2 works by running it in different contexts. Please add the following to the manual under the title DESCRIPTION, after you check my suppositions. Please remember that a newcomer who reads the manual does not know par2 in advance.
par2 protects data files by creating recovery files with .par2 extensions. It divides the data files into blocks and it creates recovery blocks saved in the recovery files. All blocks have equal sizes. The required and reported size and number of blocks are rigorously respected. Recovery blocks are ordered and you can generate any sequence of them. Their required and reported number and size are also rigorously respected. The recovery files also contain info about the data files' names and sizes. The recovery files with voln+m in their name memorize the recovery blocks n+1 to n+m. The recovery file without voln+m in its name verifies the integrity of the data files and contains no recovery blocks. The required number of recovery files is rigorously respected, but the required size of recovery files is not, probably because of the uneven number of recovery blocks. The size of recovery files is much larger than that of the recovery blocks they contain.
par2 c file - protect file, with file*.par2 as recovery files
par2 c file files - protect files, with file*.par2 as recovery files. If file ends in .par2, .par2 is removed. par2 overwrites old file*.par2 files without warning.
par2 v|r file [files] - verify / try to recover data files protected by file*.par2, fn.volp+q.par2 if file is fn.volm+n and fn.par2 if fn is in files and fn is not data file. files is a list of recovery files not identified by file and of data files which do not have their original names. If file or not data files end in .par2, .par2 is removed. files should be used in case that file*.par2's and data files' names have been corrupted. par2 is able to recover data files with data added to and deleted from any part of them. par2 is able to recover data files which have been split into separate files. par2 does not report missing file*.par2 recovery files, it reports only available recovery blocks.
These should complete OPTIONS:
-b<n> sets the number of blocks in which data files are split
-u requires equal recovery file sizes
-u is respected with approximation
-r<n> requires a level of redundancy of n% of the data files sizes. It is 5% by default.
-n<n> requires exactly n recovery files, file.par2 not included