The Rosetta 'paths.txt' file is used to tell Rosetta where to find particular
data files it needs.  Unfortunately the syntax is rather obscure and brittle.
This file tries to explain it.

A typical paths.txt file looks like this:

1   Rosetta Input/Output Paths (order essential)
2   path is first '/', './',or  '../' to next whitespace, must end with '/'
3   INPUT PATHS:
4   pdb1                            ./
5   pdb2                            ./
6   alternate data files            ../
7   fragments                       ./
8   structure dssp, ssa, dat, jones ./
9   sequence fasta, dat, jones      ./
10  constraints                     ./
11  starting structure              ./
12  data files                      ./
13  OUTPUT PATHS:
14  movie                           ./
15  pdb path                        ./
16  score                           ./
17  status                          ./
18  user                            ./
19  FRAGMENTS: (use '*****' in place of pdb name and chain)
20  2                               number of valid fragment files
21  3                               frag file 1 size
23  aa*****03_05.200_v1_3           name
23  9                               frag file 2 size
24  aa*****09_05.200_v1_3           name

The first thing to realize is that the file is (as noted on the first line)
completely order-dependent.  You cannot rearrange any of the lines,

The second thing to realize is that this file is not parsed by whitespace,
but by COLUMNS (i.e. Fortran style, not C++ style), and while the exact
amount of whitespace between the text and the values isn't precise, if you
move the values too close your paths.txt files will break.

Thirdly note that the indentation and line numbers are there for clarity
but are not in the actual paths.txt files.

Given all that the format is as follows.

    1-2, 3, 13, 19              Comments
    4-12, 14-18                 Name/value pairs.  The names are effectively
                                comments, but the values are paths and must
                                follow the path format described below.
    4:                          Primary path to find PDBs.
    5:                          Secondary path to find PDBs.       
    6:                          Secondary path to rosetta_database.  Yes the
                                secondary path comes before the primary one.
    7:                          Path to fragments
    8:                          Path to .dssp files
    9:                          Path to sequence map(s)
    10:                         Path to constraints (.cst) files
    11:                         Path to starting structures
    12:                         Primary path to rosetta_database
    14:                         Path to generated movie files
    15:                         Path to generated PDBs                      
    16:                         Path to score file
    17:                         Path to status file (scores of decoys
                                that survive a run)
    18:                         [Not sure what this is]
    20:                         Number of valid fragment files.  Twice
                                that many lines remain in the paths.txt
    21+:                        Value/name pairs.  The names are comments
                                are fragment entries, which are made up
                                of two lines:
                                - The value of the first is 3 or 9:
                                  the size in residues of the fragment.
                                - The second is the filename the fragments
                                  are in.


Those values which are paths must start with ".", "./", or "../", and must
end with "/".  It is generally best to leave these paths as relative ones
as the directories can the be retargeted via symlinks.  Leaving hardcoded
paths will cause your Rosetta runs to break whenever you move your data,
or worse run against the wrong data.

The format of the names of fragment files is explained in README.filenames.
