OPS.TXT           APRS OPERATIONS NOTES
Document version: 8.6.7 1 Dec 2003 Previous was 8.3.4 of 7 Mar 99
Author(s):        Bob Bruninga, WB4APR
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

FOR a general APRS overview see APRS.txt.
FOR MOBILE Operations, see MOBILE.txt
For operations with ALiases or MARS see ALIASES.TXT or MARS.TXT
FOR HF Operations, see HF.txt
FOR an APRS command summary see HELP.txt
FOR multi-PC operations on a piece of ZIP cord, see ZIP-LAN.txt

OVERVIEW:  This OPERATIONS file may help you to understand the finer
points of operating an APRS net both ROUTINE and SPECIAL EVENT.  Since
APRS was designed to facilitate real-time tactical communications,
operating APRS on a routine basis is sometimes like watching the grass
grow!  The reason for building a routine APRS net is primarily for
operator training and familiarity.  If your operators are not familiar
with APRS in a benign environment, then they will not be able to use it
under stress!

Although APRS is desiged for local real-time operations under stress, 
it does extend worldwide via the IGate system.  Using APRS on-line
lets you see any of the 30 thousand or so other stations.  On RF,
although you only see your local RF area, you can still keyboard with
anyone on the globe because of teh automatic linking of the IGates.
And if you are in message QSO with a distant station, the local IGate
will send you a courtesy position report of the distant station.

Again, the worldwide aspect of APRS is only there to give you something
to practice with.  This is not the design objective.  SImilarly, GPS is
a powerful tool linked to APRS.  But, do not think that you need GPS for 
tracking special events.  It is so easy to update your vehicle's position 
just by moving the cursor on your map and hitting the INPUT-MY command, 
that the only stations that need GPS are the ones that are lost!

Use APRSmax for routine FIXED station operations to see more stations.

Use APRSdos when all commands are needed and fewer stations are needed.

DIGIPEATER RULES:  Since APRS is not a connected protocol, it depends on
a fixed general path to cover a large are of operations and this provides
redundancy and repetition to assure that all stations get a copy of
everything.  But this is not an efficient process for getting message
packets from one particular station to one other station because it
floods packets everywhere.  In such cases of point-to-point messages,
ALWAYS CHOOSE a MINIMUM SPECIFIC PATH TO THAT ONE STATION.  You will
be amazed at the improvement in throughput!  Watch the DIGI page, and if
you can work him direct, DO SO!

WIDE-RELAYS AND MOBILES:  The greatest advantage of APRS is in the use of
generic alias callsigns for digipeaters so that mobiles do not have to
change their paths as they move from area to area.  Since WIDES are
widely separated and are made for WIDE area and LONGHAUL comms, many
mobiles (who have half the range due to mobile flutter) cannot hit them
reliably. For this reason, the favorite mobile path is RELAY,WIDE,WIDE
so that the mobile can be injected into the long haul WIDE hops from any
other APRS station (RELAY).  To make RELAY universal, the WIDE
digipeaters also have this alias too.  See PATHS.TXT and DIGIs.txt.
FIXED stations should NOT use RELAY generic paths except under unusual
circumstances.


ROUTINE OPERATIONS:  These days almost every area of the country has
a WIDE digipeater so the initial default path of WIDE will work.  If
a new area has no wide, then it is essential to put one up to get APRS
going in the area and to bring in distant packets.  It is important to
keep generic WIDEs well separated to minimize duplicate repeats.

ALTERNATE PATHS:  In APRSdos prior to version 860, you could store up to
12 different, frequently used DIGI paths using the OPS-DIGI command for
instant recall to tailor your DIGI path for the exact calls and path for
each QSO.  Proper use of this capability can significantly improve APRS
effeciency and reduce the use of long generic paths!  But under the new
N-N Paradigm of 2003, long paths of any kind are just not desired anymore
so this flexibility is no longer worth the significant code it used.

The following paths are reasonable under the circumstances shown:

   RELAY,WIDE  - Default.  Good starting place to see whats out there
   WIDE,WIDE   - Goes out 2 hops in all directions with minimal foldback

   WIDE,WIDE2-2 - Gets out 3 hops in all directions and identifies
                  the callsign of the original digi where the packet
                  entered the system.

The following paths are NOT considered good practice and should not
normally be used... and if you do, some Do-gooder will fuss at you...

   xxxx,RELAY      - Never put RELAY anywhere other than the first hop!
                     You may key up every station in hundreds of miles!

   WIDE,WIDE,WIDE  - Gets out 3 hops in all directions but also results
                     in as many as 27 duplicate packets in old networks
                     that have not upgraded to KPC-3 or PacComm TNCs.
                     
   WIDE,GATE       - QRMs the HF net which you cannot hear!

   WIDE,GATE,GATE,WIDE - QRMs every APRS net in the country!


WIDEn-n DIGIS:  These are APRS unique digipeters that can digipeat based
only on the number of hops indicated in the digipeater SSID digit.
For example, a WIDE3-3 packet will be digipeated out 3 hops in all
directions with little or no duplication...  See DIGIS.TXT and PATHS.TXT.

All users must understand that they are responsible for setting their
outgoing VIA path so that their packets hit the intended area of interest.
Unlike normal CONNECTED protocols which automatically return ACKS via the
reverse path of incomming packets, APRS is an unconnected broadcast
protocol only and each station's packets will only go via the outgoing
path set up by that station.  If your station receives a duplicate APRS
MSG packet more than twice, it gives you a beep and an alert that your
ACK's are probably not being heard by the other station and that you
should check your outgoing VIA path to make sure it gets to the sender.

APRS has a very useful feature for determining the best path between
stations.  The Power-Height-Gain reporting capability lets APRS plot range
contours around all stations that have included the P-H-G data in their
position reports (using the INPUT-PWR command).   Of course, Height here
is height above AVERAG TERRAIN, not above sealevel or tower height!  If  
you do not understand the difference, ASK!     See DIGIs.txt for the
format details.

Those stations between WIDE area digipeaters only need to use the single
hop of WIDE and their packets will go in both directions.  Stations that can
only hit one WIDE area station may set the path of WIDE,WIDE without any
conflicts.  Areas with TRACE or WIDEn-n digipeaters can use just about
any path without fear of duplication.  But in general it is always useful
to begin your path with your most reliable digipeater so that even without
a position report, stations can see approximately where you are by the
first digi in your path.

CAUTIONS ABOUT APRS MESSAGES:  Remember that multiple digipeater hops has
long been condemned in the packet community because the probability of
success for CONNECTED packets goes down drastically because all ACKS must
be successfully returned or packets get re-tried over and over.

This is generally NOT a problem with APRS operations which use UI
frames without acks.  HOWEVER, APRS one-line MSGS are ACKED, and the
inefficiency of multi-hops DOES APPLY!  If you do a lot of one-line
messages between operators, you will experience the same hopeless
probabilities of success as with conventional packet.  But, in general,
NEVER expect an APRS MSG to be successful beyond 2 digi's except if
everyone else is off the air!  Operator messages are a secondary
function of APRS, and should not be used as a primary means of passing
traffic!

One further caution, since APRS suspends all packet processing while
waiting for the operator with a BOXED prompt, never linger in a prompt.
The SEND command is a BOXED prompt and should not be left un-completed!

ACKS THAT DONT MAKE IT:  Just like connected packet, the chance of a
message packet getting through is usually the same as the chance that
the ACK will get back.  If the radio path is only 50%, that means that
the receiver will probably get the message by the second transmission,
but that the sender may not get an ACK until after his 4th!  This is
because the sender had to send 4 packets to get two through and the
receiver then ACKed twice in order to get one through.  You see this
effect frequently on APRS, when you are talking with a station over a
long poor path.  You will notice that the person at the other end has
already responded to your message even before you get an ack from your
outgoing message.  BUT your next line will never go out UNTIL it gets
that ACK.  The reason that you will probably get his response message
before your ack, is because his response message is being repeated over
and over in the usual APRS decayed algorithm, but his ACK is ONLY
transmitted once each time he gets a dupe of your message line to him.

What this means is that whenever it is obvious that the other station
has responded to your message line, you should ERASE it so that APRS
will move on to the next line.  Sometimes if you know that the other
station is probably hearing the digi better than the digi is hearing
him, and you are not getting ACKS, then simply send him messages in
the blind.  Let each line will be transmitted for 6 minutes and then
you can erase it.  APRS will then move on to the next line.  Remember
that APRS will have transmitted 6 times in the first 6 minutes, but
that it will then be over 3 minutes, then 6 and then 12 minutes for
further transmissions.

To improve on this effect of lost ACK responses, APRSdos recognizes a
duplicate message, and not only sends out the usual ACK, but stores a
copy for transmisssion in the blind 30 seconds later.  The 30 second
delay is to avoid cluttering up the frequency if the path is good, since
the sending station will have sent the message at least twice in the
first 30 seconds.  After the third transmission, it is clear that the
ACKs are getting lost and it is time to double up. This algorithm has
the potential of doubling throughput on a poor channel!

REPLY ACK:  In the year 2000 time frame, we figured out how to embed ACKS
in reply messages so that for only 2 more bytes in a message in a dialog
between 2 users, we can carry the ACK back for free.  This vastly improves
the reliability of ACKS in a dialog.  APRSdos and APS+SA are the only ones
that implemented it immediately, but this, combined with APRSdos decaying 
retry algorithm can result in an ORDER OF MAGNITUDE faster message dialog 
than other client software that does not use these two fundamental 
communications techniques.

SHORT MESSAGES:  As with any packet, especially on HF, the shorter the
packet the better the chance of getting through.  With 25 characters of
overhead, however, there is not much sense in making the message part
much shorter than a half line (40 characters).  The chance of a 40
character line getting clobbered compared to a 75 character line is 65%.
On HF keep 'em short.  A trick that I frequently use whenever I know that
a station is not currently on the air, or the path is not currently good,
is to send the first message line with only the word "test" followed by
additional lines with the body of the message.  This way, only the very
short "test" line is transmitted (often for hours on HF) until the band
opens, and then once the station ACKs that line, the remaining lines
are transmitted.

BULLETINS:  To send a bulletin to all stations, simply SEND a message to
BLN# where # is a line number from 1 to 9.  Like any other message, these
BULLETIN lines will be transmitted on the decaying time period and will
soon fade out of the system.  If you want the bulletin to remain at about
a 15 minute rate, then instead of using numerals in the BLN# mesage, use
a LETTER.  This way, new stations joining the net will quickly pick up
the BULLETINS.  Since lines are sorted onto the receiver's BULLETIN page,
a new BLNx line will overwrite any previous BLNx at all stations making
changes and corrections easy.  If your bulletin is time sensitive, be
sure to include the TIME in the text, since BULLETINS are not time-
stamped.  When your BULLETIN is no longer needed, simply ERASE your
outgoing BLN#.  This will stop your transmission of the BULLETIN lines.
Receiving stations can erase all old bulletins by using the ALT-E command.


GATEWAY RULES:  I have interjected this paragraph because of the large
number of APRS HF to VHF gateways now in operation.  First, it is very
important that users understand that GATEWAYS ARE ONLY INTENDED TO LINK
HF ACTIVITY INTO LOCAL VHF NETS.  IT IS INNEFFICIENT, DISCOURTEOUS, AND
MAYBE ILLEGAL TO LINK from VHF to HF.  Linking HF operations into every
local VHF APRS net in the country is not a problem, because the slow 300
baud data rate could never saturate ANY 1200 baud local net.  HOWEVER,
linking just ONE active VHF net ANYWHERE in the country out onto HF
WOULD CERTAINLY BLOCK ALL HF OPERATIONS NATIONWIDE!  The capability is
there for linking special events or cross country travelers on VHF out
for the entertainment of all HF listeners, but DO NOT ABUSE IT, OR WE
WILL LOOSE IT!  See HF.txt.

GATE SUMMARY:  On HF, use the path of GATE,WIDE and everyone in the
country within one WIDE hop of a GATE will likely see you.  *Never*
use GATE,WIDE,WIDE because your packets will now go 2 hops on VHF and
be seen MULTIPLE times from multiple gates!  No one can tell where the
GATE is and it is just a BIG MESS.  Believe me!

Second, never routinely go through a GATE on VHF.


USING THE OPS-COMM-TNC dumb terminal mode.  This mode works OK for using
your TNC normally, but it doesnt have any file transfer capabilities.  You
might try to find a small .EXE comm program that you can FILES-SHELL out
of APRS, do your COMM thing, and then EXIT back into APRS....  YAPP and
PACCTERM both work, but be careful of how these programs change your
TNC parameters...

OBJECTS:  As noted previously, anyone may place an object on the map and
all other stations will see it.  On their P-list, the object will be
marked with the last three letters of the originating station.  Any other
station that has more current information on that object can also update
its position by SELECTing, moving the cursor, and then hitting the insert
key.  His station will begin uplinking the new posit, and all stations,
will update their P-list entry for that object INCLUDING THE ORIGINAL
UPLINK STATION!  Since the new position overwrites the old one, the
original originating station will now no longer uplink it.  This comes
in handy during hurricane tracking.  Who ever has information on the
latest HURICANE posit uplinks it and everyone then always sees the latest
storm track without anyone in the net being dependent on any one station
for updates!

    Once objects are transmitted on to other station map screens, they will
remain there until that operator deletes them, even if the originator stops
transmitting them.  It will, however, fade to dark gray after 2 hours to
show it as an old report.  You can use the CONTROLS-FADE command to bring
them back to bright colors, or use the J command to see JUST-the-LATEST
symbols.  The KILL function permits the originator of an OBJECT KILL it
from all displays on the net.  His station will continue to uplink the
object, but tagged with a special KILL flag to suppress its display on all
screens.  It remains in everyone's P-lists, though, so they can refer back
to it if needed.  They must still manually DELete it from their P-list as
needed.  Once the originator has KILLED an object, he should let it remain
on his P-list for at least 6 minutes to be sure everyone has received the
KILL indicator; then he can delete it from his list.


NEAT OPERATOR FEATURES:

There are several  menus and commands you can use to set up your operating
envrionment.  Here are some highlights:

CONTROLS MENU:  You can set filters and how you want packets displayed.
MAPS:           Turn on or off most map features plus overlay data files
                on the maps such as DIGIS, Radio Shacks, etc
RADAR ALARMS:   Use INPUT-MY-RADAR to set your"airspace".
POSIT ALARMS:   Set ALARMS on any mobile which will alarm if he moves.
WX ALARMS:      Set Weather alarms on temps, winds or rain...
TRACK MODE:     Lock on to one station and keep map centered
SPECIAL MARKS:  Mark any station on the P-LIST for SPECIAL highlighting
                Then you can see only those stations with JUST-S



SPECIAL EVENT OPERATIONS:

    The alt-SETUP-MODES-SPECIAL command sets up an APRS station to send
TO the UNPROTO address of SPCL... vice APRS...  and to ignore all other
packets NOT addressed to SPCL.  This allows the event participants to
keep their screens (P/L lists, etc) clear of unwanted other APRS stations
on frequency, while tracking the event normally.  All other stations
watching the event will still receive all SPCL event posits on their
screens, and they will be automatically marked with the # for special
display using the JUST-SPECIAL command or SPACE bar.

SPECIAL EVENTS:  The Cycle Across Maryland (CAM) bike tour is a good
example of a special event using APRS.  We had two of three relief
vehicles with GPS trackers.  These were assembled in cake pan enclosures
duct-taped to the roof with a small power cable extended down the
windshield and clipped directly to the battery.  These packages could
be moved among vehicles in about five minutes.  Some other packet
mobiles ran APRS without GPS units by just using the INPUT-MyPOS command
to update their positions.

Since we only have two WIDE digipeaters in the state, we were dependent
on many home RELAY stations all across the state to serve as digipeaters
for the event.  We simply asked many conventional packeteers to tune to
144.39 for the event and set their TNC aliases to RELAY for that day.

We also set up both GPS units with the alias of RELAY so that they
would also help digipreat each other along the trail.  The disdavantage
of this technique however, was evident as both vehicles returned to
the evenings command post (also RELAY) and you had three RELAYS in 100
yards of each other!

SYMBOLS:  APRS now permits HUNDREDS of different mobile symbols by using 
the NUMERIC OVERLAY capability.  This makes it easy to distinguish
mobiles even with CALLSIGNS-OFF to reduce clutter!

EMISSION CONTROL:  If there are only a few APRS stations involved in an
event but there are lots of APRS observers on frequency, then the observers
can set their transmitter off using the CONTROLS-X command to minimize QRM
on channel.  They can still transmit under manual control by using the
X key.


LOAD SHARING:  Since any station can take over reporting of any objects,
one approach is to let only one station SELECT every symbol that comes in
and then he becomes the reporting repsonsibility.  The original station
that uplinked the report in the first place will fall silent when it sees
the report comming from the designated Net Control station.  This way all
positions are reported by only one station on frequency, although all
other stations can still update the positions as needed.  Remember that
the last station to report the position of an object will be the one
that continues to report it!  APRSdos has a NET-CONTROL feature for
OBJECTS that automates this process.


MARINE CORPS MARATHON:  See MARATHON.txt for the lessons learned using
APRS at the Marine Corps Marathon for the last 3 years in Washington DC.


ZIP_LAN MODE AND EMERGENCY OPERATIONS CENTERS:  Dont overlook, that a
handful of separate PC computers can ALL BE CONNECTED TO A SINGLE TNC AND
RADIO!  This fact can be used to create quite an impressive multi-station
tactcal communications system that will rival some 911 consoles! No
special LAN hardware is required other than a serial port and as much
two conductor zip cord as you need.  See ZIP-LAN.txt

CAUTION:  This ZIP-LAN capability is not backwards compatible to any
software prior to APRS800, Mac/Win prior to 2.09 and APRSa4 ver 0408.

With a ZIP-LAN, ALL consoles see the tactical picture, and these PC's
are at the individual operator's disposal to zoom in, and hop from screen
to screen to give them access to what ever info they need!  Do not think
that a big screen display is better.  A single big screen is impressive,
but actually useless.  Only the person at the KEYBOARD of an APRS system
can actually get useful info from APRS.  In our county, you need to be
below the 8 mile scale to get an idea of what is going on at a crisis,
and while you are zoomed in there, others need to be focusing on other
parts of the county, or different screens.

You can wire every PC in the building using cheap 2 conductor speaker
ZIP cord!  You can carry hundreds of feet of this stuff in your briefcase
with your portable laptop!

This is a TREMENDOUS capability, since these days PC's are much more
plentiful than TNC's and all available assets can be brought into the
picture.  Every SLAVE operator has his own INDEPENDENT access to all
of the APRS info without bothering the APRS operator.

See ZIP-LAN.TXT

de WB4APR, Bob
