* THROUGHPUT VS JITTER graph shows throughput (bits) on X axis and jitter on Y axis. It's similar to Throughput vs delay graphs.
  If S(i) is the time packet i was sent from the sender, and R(i) is the time it was received by the receiver, 
  the jitter(i)=|(R(i+1)-S(i+1))-(R(i)-S(i)|. 
* THROUGHPUT VS PROCESSING TIME graph shows throughput (bits) on X axis and processing time on Y axis.
  Throughput and processing times (minimal/average/maximal) are calculated every TIME INTERVAL. Throughput of receiving/forwarding bits 
  and processing times are calculated at Current node or in the whole network (simulation processing time).
* THROUGHPUT VS RTT graph shows throughput (bits) on X axis and RTT on Y axis.
  Throughput and RTT (minimal/average/maximal) are calculated every TIME INTERVAL. Throughput of sending bits 
  is calculated at Current node or in the whole network. RTT is calculated between Current node (source) and Other node (destination) or
  in the whole network between all the sources and destinations. See 2.9, 2.10.
* PACKET SIZE VS THROUGHPUT graph shows packet size on X axis and minimal/average/maximal throughput on Y axis.
  Throughput is calculated in each TIME INTERVAL and then minimal/average/maximal value is calculated from all the time intervals.
* PACKET SIZE VS DELAY graph shows minimal/average/maximal delay in the whole network (simulation End2End delay) or delay between
  Current and Other node. See Direct Connection option.
* PACKET SIZE VS JITTER graph shows minimal/average/maximal jitter in the whole network (simulation jitter) or jitter between
  Current and Other node. 
* PACKET SIZE VS PROCESSING TIME graph shows minimal/average/maximal processing times in the whole network (simulation processing time) 
  or processing times between Current and Other node.
* PACKET SIZE VS RTT graph shows minimal/average/maximal RTT in the whole network (simulation RTT) or RTT between
  Current and Other node. See 2.9, 2.10.
* JITTER (of ...) (Y axis) = current event (generate/send/receive/forward/drop) time -
  previous event time /  current event sequence number (X axis) - previous event sequence number.
* SEND EVENTS TIME vs JITTER graph shows time when packets were sent (X axis) vs packets jitter (Y axis).
  If S(i) is the time packet i was sent from the sender, and R(i) is the time it was received by the receiver, 
  the jitter(i)=|(R(i+1)-S(i+1))-(R(i)-S(i)|. Jitter definition is the same for RECEIVE EVENTS TIME vs JITTER graphs
  and for PACKETS IDs vs JITTER graphs. SOURCE AND DESTINATION TRACE LEVEL options are used for the jitter calculations.
* RECEIVE EVENTS TIME vs JITTER graph shows time when packets were received (X axis) vs packets jitter (Y axis).
  SOURCE AND DESTINATION TRACE LEVEL options are used for the jitter calculations.
* PACKETS IDs vs JITTER graph shows sent packets IDs (X axis) vs packets jitter (Y axis).
  SOURCE AND DESTINATION TRACE LEVEL options are used for the jitter calculations.
* SENT PACKET ID vs RTT graph shows sent packets IDs (X axis, not ack packets) at the source vs RTT (Y axis). See 2.9, 2.10.
* SEND EVENTS TIME vs DELAYS graph shows time when packets were sent (X axis) vs packets delays (Y axis).
* RECEIVE EVENTS TIME vs DELAYS graph shows time when packets were received (X axis) vs packets delays (Y axis).
* FORWARD/RECEIVE EVENTS TIME vs PROCESSING TIMES graph shows time when packets were forwarded/received (X axis) 
  vs processing times (Y axis).
* CUMULATIVE DISTRIBUTION is an empirical cumulative distribution function.
* All send events time vs number of intermediate nodes graph shows time when packets were sent (X axis, for the whole
  network) vs number of intermediate nodes (Y axis) forwarding (fwd) or receiving (rec) the packets between the source node 
  and the destination node. All receive events time vs number of intermediate nodes graph is similar. DIRECT CONNECTION MUST BE OFF.
* Send events time at CN vs number of intermediate nodes graph shows time when packets were sent by CURRENT NODE (X axis)
  vs number of intermediate nodes (Y axis) forwarding (fwd) or receiving (rec) the packets between the source node (CURRENT NODE) 
  and the destination node (OTHER NODE). Receive events time at ON vs number of intermediate nodes graph is similar. 
  DIRECT CONNECTION MUST BE OFF.
* All packets IDs vs number of intermediate nodes graph shows packets IDs (X axis, for the whole network) 
  vs number of intermediate nodes (Y axis) forwarding (fwd) or receiving (rec) the packets between the source node and the destination node.
  Packets IDs vs number of intermediate nodes concerns only CURRENT NODE (the source) AND OTHER NODE (the destination). 
  DIRECT CONNECTION MUST BE OFF.
* For PACKET ID vs DELAY, PACKET ID vs PROCESSING TIME, PACKET ID vs JITTER, PACKET ID vs RTT and number of intermediate nodes graphs 
  packets IDs are sorted in ascending order.
SOURCE AND DESTINATION TRACE LEVEL options are used for all the End2End delays graphs.
See 2.3 and 2.4 for DELAYS details.
See 2.7 and 2.8 for PROCESSING TIMES details.

5. 3D GRAPHS
============

* 3D graphs can be rotated. To rotate 3D graphs hold left mouse button.
* CN means source node (X axis on the right side) and ON destination node (Y axis on the left side, see 8. Shortcuts, 3.4).
* When there are no packets transmitted an empty graph is plotted (Z equals zero).
* Sometimes when DIRECT CONNECTION option is ON (see 3.4) SUM OF NUMBER OF LOST PACKETS AT ALL THE NODES 
  graph can show higher values when PACKET TYPE option (see 7) is ON than if it's OFF. It means that 
  CURRENT NODE forwarded packet(s) directly to OTHER NODE and forwarded packet(s) with '-Id' field = -1 
  (see 2.1, 9.6) or forwarded packet(s) only with '-Id' field = -1. 
  OTHER NODE received some of them (if lost) both directly forwared and with '-Id' field = -1.
See 3.5 for details about some 3D graphs.


6. HISTOGRAMS
=============

* histograms show frequency distribution
* Each histogram has N containers where N=NUMBER OF HISTOGRAM INTERVALS.
* Delays histograms show delays on X axis and number of packets (delays) on Y axis.
* Delays between 2 nodes are calculated in the same way like in 2.4.
* Processing times histograms show processing times on X axis and number of packets (processing times) on Y axis.
* When USE SPECIFIED EDGES FOR HISTOGRAMS option is off NUMBER OF HISTOGRAM INTERVALS (containers) is used.
  In the other case specified egdes (in the field HIST. EDGES) are used. Only values inside the edges are counted, 
  e.g. for edges 0.03 0.05 0.07 values are counted between 0.03 and 0.05 (>=0.03 and <0.05), between 0.05 and 0.07 (>=0.05 and <0.07), 
  and the last container shows count of any values of X that match the last value of HIST. EDGES (=0.07).
  To plot value count e.g. greater than 0.05 (>=0.05) type 0.05 inf in the HIST. EDGES field.
  To plot value count e.g. lower than 0.05 (<0.05) type 0 0.05 in the HIST. EDGES field.


7. OTHER OPTIONS
================

<-> button                     - switches between Current and Other node
CURRENT NODE                   - current node number
END TIME                       - time interval end (see 3.3)
FLOW ID                        - flow identificator (see 3.15 for details)
HIST. EDGES                    - histogram edges (see 3.7 for details)
NUMBER OF HISTOGRAM INTERVALS  - number of histogram containers
OTHER NODE                     - other node number
PACKET TYPE                    - enables calculations for selected packet types (selected: +, deselected: -)
                                 to select/deselect a packet type choose it from the list
                                 all the types are deselected by default
                                 if this options is not checked all the types are used in calculations
PACKET SIZE                    - enables calculations for selected packet sizes (selected: +, deselected: -)
                                 to select/deselect a packet size choose it from the list
                                 all the types are deselected by default
                                 if this options is not checked all the sizes are used in calculations
SELECT/DESELECT ALL SIZES      - selects or deselects all packets sizes from PACKET SIZE list
SOURCE/DESTINATION TRACE LEVEL - trace level at source/destination node used for End2End delays and some jitter calculations 
                                 (see 2.3, 2.4, 4)
START TIME                     - time interval beginning (see 3.3)
TIME INTERVAL LENGTH           - for throughput graphs (see 4. 2D Graphs)
TITLE                          - sets graph title
X, Y, Z AXIS LABEL             - sets labels for the axes


8. SHORTCUTS
============

Shortcuts are used in network information menu or graphs legend.
ACK:        <acknowledge packet type> (for RTT)
CN:         <current node number>
DC:         <direct connection>
DTL:        <destination trace level>
ET:         <end time>
FID:        <flow ID>
HISTEDGES:  <histogram edges>
HISTINT:    <number of histogram intervals>
IDx1:       <count packets IDs only once is on> (see 3.5)
LSX:        <log scaling on X axis>
LSY:        <log scaling on Y axis>
LSZ:        <log scaling on Z axis>
MARKPOINTS: <marks plotted data points with circles>
MAJORGRID:  <adds major grid lines to the axes>
MINORGRID:  <adds minor grid lines to the axes>
N/A:        <not available> (see 2.4)
node:       <node number>
ON:         <other node number>
PID:        <packet ID>
PT:         <packet type>
PS:         <packet size>
[sec]:      <seconds>
seq#:       <sequence number>
SPAK:       <sent packet type> (for RTT)
SPID:       <sent packet ID> (for RTT)
ST:         <start time>
STL:        <source trace level>
TIL:        <time interval length> (for throughput graphs)
TM:         <turbo mode>
X:          <description of X axis>
Y:          <description of Y axis>


9. SUPPORTED FILE FORMATS 
=========================

* Trace graph supports all NS trace file formats (one wired, two wireless, two wired+wireless) except http trace format.
  Examples of these formats are shown below. Lines are copied from some files randomly.
  Trace graph also has it's own trace file format. You can convert your own trace files into Trace graph 
  file format if your trace files are different from formats described in 9.1, 9.2, 9.3, 9.4, 9.5. See chapter 9.6 for details.
* Trace graph reads trace files in wireless format (see 9.2) with HEXADECIMAL nodes numbers in the 11th column
  ('-Ms' field in new trace file format, see 9.3, 9.6). It converts all the numbers to decimal
  and additionally creates 'temp' file in Trace graph format (see 9.6) in case some of the numbers can
  be decimal before conversion. You can change these numbers in the 8th column (see 9.6) in 'temp' file yourself.
  For example, a line before conversion:

  s 5.000075000 _10_ MAC  --- 0 AODV 104 [0 ffffffff a 800] ------- [10:255 -1:255 1 0] [0x2 0 1 [11 0] [10 1]] (REQUEST)

  and after hex to dec conversion (a=10, 8th column):

  s 5.000075000 10 MAC 0 AODV 104 10 10 -1 0 -1

  NOTE! Header 'hex' has to be added to a trace file in format 9.2 to turn hex to dec conversion on. For example:
  
  hex
  M 0.0 nn 20 x 670 y 670 rp AODV

* Files with IP addresses like e.g. 1.2.3.0 or 4123567:1 must have 'ip' header (doesn't concern Trace graph format, because it has no IPs),
  for example:
  
  ip
  + 1 1 0 cbr 1000 ------- 2 0.1.0.2 1.0.1.2 0 0

  All the IPs must be converted into nodes numbers before loading a trace file.
  To get all the IPs from a trace file you can type '-ip' after file name, e.g.
  trgraph c:\out.tr -ip
  You can also get IPs within a time interval (see 1.1, 3.6), e.g.
  trgraph c:\out.tr -ip 1.1 2.3
  IPs are saved to a file which name has the following format: trace file name + '.ip' extension.
  If time interval is specified the file name format is e.g.: trace file name + 'ST_1.1' + 'ET_2.3' + '.ip' extension (see 8).
  If a trace file is loaded from the File menu the file with IPs is generated automatically if doesn't exist.
  You have to define relations between nodes numbers and IPs in the saved file with IPs. Below we have an example of such a file:

  0.1.0.2 1:2
  0.3.0.1 0:1
  1.0.1.2 2:2
  1.0.3.1 3:1
  2048:2 5:2
  4194305 6
  4194305:2 6:2
  4194307:1 4:1
  6144:1 7:1

  You've got to create the second column with node and (optionally) port number.
  For wireless format (see 9.2) and wired + wireless format (see 9.4) there have to be ':' between node and port number.
  For wired format (see 9.1), new trace file format (see 9.3) and wired + new trace file format (see 9.5) there have to be '.' 
  between node and port number. When IP file is created properly trace file can be loaded. Trace graph converts all the IPs into
  nodes numbers line after line (strings are replaced) and writes an output trace file which name format is: trace file format 
  without extension + 'no_ip.tr'. When you have the file with IPs you can do the conversion in another way - it can be much faster.
  Then trace file with changed IPs is converted to Trace graph format, saved in 'temp' file and loaded.
  NOTE! Because IPs are converted in the order they appear in the file (with '.ip' extension) sometimes you've got to remove some IPs 
        or change their order to avoid replacing other strings in a line or already replaced IPs. 
  For example first line must be removed because it's not neccessary:
  1:0