OpenBCM V1.13 (Linux)

Packet Radio Mailbox

DB0FHN

[JN59NK Nuernberg]

 Login: GUEST





  
ZL3AI  > APRDIG   04.02.03 10:08l 298 Lines 12847 Bytes #999 (0) @ WW
BID : 833-ZL3AI
Read: GUEST
Subj: TAPR Digest, Jan 30, 7/8
Path: DB0FHN<DB0ZWI<DB0HST<HB9OK<IR0NET<VK3TE<ZL2BAU<ZL3VML
Sent: 030203/0922Z @:ZL3VML.#80.NZL.OC #:58726 [Chch-NZ] FBB7.00i $:833-ZL3AI
From: ZL3AI@ZL3VML.#80.NZL.OC
To  : APRDIG@WW

Subject: RE: New spec - working notes
From:     Jeff King <jeff@aerodata.net>
Date: Thu, 30 Jan 2003 19:21:11 -0500
X-Message-Number: 50

>>>The big problem with APRS is total overload on the national user channel,
>>>and I dont see that any of that is a significant "protocol" problem.
>>
>>What?!? Lets not continue this group psychosis. There are *serious* flaws in
>>the level one and level two implementations of APRS, and these are a
>>fundamental part of the protocol.
>
>I'd be interested in knowing what they are so we are all working from the
>same baseline..

Sure, take a look at this page for a good description of the problem:

http://web.usna.navy.mil/~bruninga/aprs/fix14439.html

as long as you temper this author's opinions with some of the serious
text's on packet radio, it should give you a good idea of the flaws and
oversights in implementation.

Understand that layer one and two of the protocol is the radio and access
layer. Possibly you didn't understand this in replying, but it is one in
the same.

>>More to the point, lets compare two *micro-controller friendly* protocols
>>that give the positional accuracy in the meter or less range....on the
>>other hand the base 2 protocol Scott is suggesting around 18 bytes. At level
>>two of the protocol, that is about a 75% decrease in packet size.
>
>Hummh... my APRS Mic-E format puts these items in a PIC-friendly format
>and only takes 9 bytes and is used in every Kenwood and Mic-E type device:
>
> CALL, LAT/LONG, CSE, SPEED, ICON, and one-of-7 messages.

60 foot granularity though, besides the well documented mangling issues.

>If you want 1 foot accuracy, then use my APRS Compressed format that can
>put all of the same info into 14 bytes...

No trackers, let alone a micro-controller tracker, currently implement this.
BASE-91 is not micro controller friendly.

>Again, I have nothing against inovation, just clairfying that a new
>protocol will have to get down to a payload smaller than 9 bytes before it
>can claim to be more "efficient"...

Not true. The "new protocol" only need be implemented before it can claim
to be more efficient. Correctly me if I am wrong, but there are no tracker
implementations of compressed format, for the very reasons I described
above.

I imagine it could be made to fit as micro's keep getting bigger and
bigger, but why continue to cling to it since no-one has implemented it?
Give it the well deserved burial it deserves and replace it with a binary
precision format that is easy to parse and doesn't require floating point.

----------------------------------------------------------------------

Subject: RE: New spec - working notes
From: Miller Scott Contr 30SCS/FTI <Scott.Miller@vandenberg.af.mil>
Date: Fri, 31 Jan 2003 00:25:29 -0000
X-Message-Number: 51

I can't blame Bob too much for the Mic-E mess - I probably would have come
up with about the same thing if I sat down and designed a position reporting
message to fit in the absolute minimum space in a standard AX-25 message
with (almost) only printable ASCII.  What really bugs me is this:

KD7ODD-1>SVPXUY,N7ZEV-2*,WIDE,qAS,N7TTE-10:'+*)l"Ck/]";{}Ray in Vegas.

Not to pick on Ray in Vegas, but without that 13 characters of comments, the
position could have been stored in a much simpler format, with better
resolution.  Take a look at the APRS-IS feed and see how many Mic-E or
compressed position reports have only the minimum possible length.  From the
quick sample I took, it's a distinct minority.

Scott / N1VG

----------------------------------------------------------------------

Subject: RE: New spec - working notes
From:     Jeff King <jeff@aerodata.net>
Date: Thu, 30 Jan 2003 19:43:06 -0500
X-Message-Number: 52

On Fri, 31 Jan 2003 00:25:29 -0000, Miller Scott Contr 30SCS/FTI wrote:
>I can't blame Bob too much for the Mic-E mess - I probably would have come
>up with about the same thing if I sat down and designed a position reporting
>message to fit in the absolute minimum space in a standard AX-25 message
>with (almost) only printable ASCII. 

I wasn't. If anything, given the conditions he had to work with back in the
mid 1990's, he came up with a silk purse out of a sow's ear. I'm just
suggesting with a "clean slate" new protocol we might want to not again
use sow's ear's, so to speak, as our base material, that is all.

----------------------------------------------------------------------

Subject: Unproto
<7BF8A28ADD2DD511B5020090278586EE03C61903@fsxumu03.vandenberg.af.mil>
From: kb2scs@optonline.net
Date: Thu, 30 Jan 2003 19:48:31 -0500
X-Message-Number: 53

Hi Scott

Ok now I am starting to understand what you are doing.

In answer to your question. The reason why UI frames are used in APRS is
because in the very beginning when Bob came up with APRS he designed APRS
to use a TNC in converse mode. When a TNC is in converse mode it sends UI
frames.

Unproto is not a frame.  Rather it is a TNC command that is used to set up 
the destination call sign and the digi path to reach that destination. 

I left out a few other bytes in the AX25 UI frame in my earlier email.

I am not sure but I think a TNC in Kiss mode also expects a control byte
and a pid byte.

A UI frame has a control byte of 03 hex which means that the packet is an
UI frame.

A UI frame has a Protocol IDentifier byte of F0 hex. This means that there
is no higher level protocol.

Not sure if this answered your question or not.

----------------------------------------------------------------------

Subject: RE: New spec - working notes
From: Steve Dimse <k4hg@tapr.org>
Date: Thu, 30 Jan 2003 20:21:22 -0500
X-Message-Number: 54

On 1/30/03 at 4:08 PM hacker@tc.fluke.com (Curt Mills, WE7U) sent:

>>Yeap. Does it fit on  1K PIC controller?  (no) What trackers presently
>>support the compressed format "BASE-91" implementation? (none) What client
>>programs support this?
>
>Xastir for one.  Perhaps the only one.

FindU and UIView do for sure, dunno about others.

Steve K4HG

----------------------------------------------------------------------

Subject: RE: New spec - working notes
From: "Ramon F. Kolb" <ramon@ieee.org>
Date: Thu, 30 Jan 2003 20:31:46 -0500
X-Message-Number: 55

Right, and this is why the "Ray in Vegas" part is OPTIONAL. This is why
in most protocols the following order is maintained:
1) mandatory, fixed-length parameters ("V", just the values) in a fixed
order
2) mandatory variable length parameters ("LV", length followed by value
for each parameter), in a fixed order
3) the optional, fixed-length ones ("TV" Tag followed by Value)
4) the optional, variable-length ones ("TLV" Tag, Length, Value)

Of course 3) and 4) can occur in any order...

This way you can make a minimal transmission by using only the mandatory
parameters, and add to it if needed. This is what both Mic-E implement
and "standard" APRS, which the difference that in "standard" APRS the
"L" component is replaced by an "end-of-argument" character, the "/"
symbol. MicE (I could be wrong as I haven't studied this too closely)
uses a similar combination with mandatory parameters upfront, followed
by optional ones. The advantage of MicE is that it's not text-based and
therefore shorter. The advantage os "standard" APRS is, that its
parameters are immediately recognizable in a terminal window. But I saw
a different thread about this, so let's not start all over again :)

Last, let's PLEAZZZZZZZ stop using Feet, Miles, Fahrenheit, etc. This
confuses the heck out of any non-US based person, HAM or not... After 7
years in this country, I still have no clue how far a highway exit at
1500ft actually is.... Eh.... 1500 ft divided by 3 is about 500 meters,
right? By the time I've done that, I passed the exit :)

--Ramon

Ramon F. Kolb, W1/PA3EUG, Boston, MA, USA [FN42KI] ICQ:77117744;
MSNMGR:legalalien30@hotmail.com http://kolb.webhop.org

----------------------------------------------------------------------

Subject: RE: Input needed
From: "Ramon F. Kolb" <ramon@ieee.org>
Date: Thu, 30 Jan 2003 20:32:09 -0500
X-Message-Number: 56

Just a question... If we use a different PID in AX.25, will this affect
the "digipeatability" of the frame by (especially) "dumb" digi's, such
as simple TNC's? I suspect that this is NOT the case, if I remember from
the "old" (1980s) days that NET/ROM, IP, etc. frames are as well
digipeated as a "text" frame. Right?

--Ramon

Ramon F. Kolb, W1/PA3EUG, Boston, MA, USA [FN42KI] ICQ:77117744;
MSNMGR:legalalien30@hotmail.com http://kolb.webhop.org

----------------------------------------------------------------------

Subject: RE: Input needed
From: "Ramon F. Kolb" <ramon@ieee.org>
Date: Thu, 30 Jan 2003 20:39:21 -0500
X-Message-Number: 57

Hi Scott and all others,

here's my personal preference:

Unproto/Layer3: if this is an APRS-IS effort only, then don't we use IP
on layer 3? If it's over the air, and we want to continue using "packet
radio" based systems (which I'd strongly recommend because of
ubiquitusness and backward compatibility), I think that we need to
continue to use broadcast-style packets only, that have as little
redundancy in them as possible. Encapsulating TCP/IP over AX.25 is not
really efficient, and makes transmissions longer than necessary

Station speed: I don't think there's a lot of difference between using
m/s or km/h. However, if you want to keep it strictly "basic" SI, go for
m/s as this is the official SI unit

Wind speed: To avoid confusion, I'd say use the same unit as for the
station speed. However, for both, make sure that their range and
accuracy is broad enough. This means, that it should have the ability to
cope with speeds in (for example) 0.01 m/s increments, and up to at
least a multiple of "satellite" speed, or a multiple of the fastest
object that we can foresee being used now.

As for temperature, I'm open for suggestions. However, <BOLD STATEMENT>
I think that anyone claiming to be able to measure ambient temperature
at better than 0.1 deg C accuracy, is full of.... Eh... Well... You
know... </BOLD STATEMENT>. Again, if you are "going SI", use Kelvin
instead of Celsius/Fahrenheit/Reumur/whatever.

Maidenhead: Yes, leave it in. I know several European, inner-city APRS
users that don't really want to put their home on the map. They use PO
Boxes for their mailing addresses, and are generally afraid that
pointing their home on a map is an invitation for burglars to come in.
Whether that fear is justified or not is for another discussion at some
other time, but I wouldn't want them to be no longer part of the APRS
world just because of that. Other APRS applications, such as VHF DX
spots which are only known by [XX##YY] locator, and BEACONet
applications (although not strictly APRS and probably not part of the
APRS-IS) rely on the maidenhead system rather than the regular system.
I'd actually be in favor of broader support for alternative locator
systems and datum sets. Every so often, I see a discussion coming up in
the UI-View mailing list about which datum to use, since the "standard
datum" for e.g., the UK, is different from the one that we use here in
the US.

I'd really encourage you to float the drafts of the new spec to as many
international (=non-USA) HAMs as possible. Nonewithstanding the effort
and adoption rates that is being put into this by US Hams, a major
complaint about the existing APRS spec is that it's way too US-centered.
If you want to make this into a truly worldwide spec, please solicit the
comments from ARPS and Packet groups around the world. This would
include the "usual suspects" as in the Europeans, Aussies/Kiwis, etc.
but we should also try to get acceptance for this in Japan. APRS is not
really well-adopted there, and their participation in a new
specification could be a good first step to get them on-board.

Lastly, professionally I used to be a Systems Engineer, writing
technical specification for voicemail systems, pretty much writing the
same type of specs you are now writing. I'd be happy to review and
comment on your drafts. In my experience, the best way of getting people
constructively involved is to write a draft, send it out for review
(even if it's not very good), take comments and implement them to your
best judgement (since people's opinions often conflict on the same
subject), and produce a new draft clearly indicating what has changed.
You can do this, for example, in MSWord by using the "tracking" feature.
If you need decisions on "conflictive" issues, have a small group of
people who are "most representative" hash it out. We need a clear
"owner" of the effort...

Best 73,

--Ramon

Ramon F. Kolb, W1/PA3EUG, Boston, MA, USA [FN42KI] ICQ:77117744;
MSNMGR:legalalien30@hotmail.com http://kolb.webhop.org

----------------------------------------------------------------------



Read previous mail | Read next mail


 16.11.2026 06:30:45lGo back Go up