| |
ZL3AI > APRDIG 07.02.03 12:00l 257 Lines 10296 Bytes #999 (0) @ WW
BID : 854-ZL3AI
Read: GUEST
Subj: TAPR Digest, Feb 03, 4/6
Path: DB0FHN<DB0RGB<OE5XBL<OE3XSR<OK0PBX<OK0PRG<OK0NAG<OK0PPL<RZ6HXA<IK1ZNW<
IR0NET<VK3TE<ZL2BAU<ZL3VML
Sent: 030205/0851Z @:ZL3VML.#80.NZL.OC #:58835 [Chch-NZ] FBB7.00i $:854-ZL3AI
From: ZL3AI@ZL3VML.#80.NZL.OC
To : APRDIG@WW
Subject: Re: [ Miller Scott Contr 30SCS/FTI ] Re: Future spec
From: "Gregg G. Wonderly" <gregg@skymaster.cytetech.com>
Date: Mon, 03 Feb 2003 12:43:56 -0600
X-Message-Number: 17
>I think the way I've seen this dealt with other places is by the use of
>DEFINEs to alias the appropriate data type.
>
>Do you have any suggestions on how the format could be structured to better
>allow cross-platform decoding?
When you use #define, then every platform has to be accounted for, and in
some designs, subtle problems are introduced because of the in ablility to
be compatible with the design.
The only way to do this is to specify that the values are 16bit, signed
values, or 32 bit signed values that are in MSB order, and go no futher. A
C library to read the data stream would be designed with functions such as
the following.
// Longitude is encoded as degrees, minutes, seconds, and 1000's of seconds.
// the format is 2bytes, 1 byte, 1byte, 2bytes
int getLongitude( int idx, char *buf, int size, float *lon ) {
if( idx + 6 >= size )
error( "buffer space too short for longitude");
float val = 0.0;
// Align sign extension for negative degrees on MSB.
int deg = (buf[0]<<8) | (buf[1]&0xff);
int min = (buf[2]);
int sec = (buf[3]);
int tsec = (buf[4]<<8) | (buf[5]&0xff);
*lon = (deg*3600)+(deg*60)+sec+(tsec/1000.0);
return idx+6; //point to next byte
}
But, in the end, you still have portability issues with the use of 'int'.
It really is a giant ugly ball of goo to write a library for use on small,
unlimited processors that have a wide choice of representations. Binary
data manipulation is one of the most problematic areas of C portability.
All you have to do is look at the history of type management.
#ifdef m68k
#define USHORT unsigned short
#define ULONG unsigned long
#endif
#ifdef m32
#define USHORT long
// Use the platforms 64 bits for unsigned long
#define ULONG long long
#endif
This kind of stuff results in code that is very hard to understand, and
even harder to maintain when the size of things change, such as pointers
that someone stuffed into an integer because it was convienent...
This alone is one of the primary reasons that I use Java for as many things
as I can. In Java, everything is specified... Much, much, much smaller
chances for portability problems that are not just implementation bugs that
the vendor is responsible for...
Don't get me wrong, I think there is a time and a place for binary data.
But, as a general example of lessons learned, look at enterprise data
messaging and brokering systems and see how the gravitation to XML
formatted data has happened so quickly. Removing data interpretation as a
potential cause of problems is a dramatic simplification.
Java enjoys platform independent binary data exchange because everything is
fully specified. All values are signed, and all sizes are fixed. There
are not pointers to have mismatched sizes with integer-valued
representations.
-----
gregg@cytetech.com (Cyte Technologies Inc)
----------------------------------------------------------------------
Subject: Capabilities announcement
From: Miller Scott Contr 30SCS/FTI <Scott.Miller@vandenberg.af.mil>
Date: Mon, 3 Feb 2003 18:46:00 -0000
X-Message-Number: 18
I'm working on an expanded station capabilities announcement to indicate
the following information about available frequencies and modes:
Frequency
Mode (probably derived from WARC-79 emission designation)
Function
Uplink/Input
Downlink/Output
Beacon
Simplex
Bandwidth (either signal bandwidth or bandpass transponder bandwidth)
I'm sure I'm missing some here, so let me know what else needs to be
announced. Repeaters, BBSs, and satellites should be able to clearly
convey their capabilities this way. Ought to help the BEACONet folks, too.
Uplink and downlink frequencies will need to be associated with each other
in the case of, for example, a satellite with multiple transponders. Is a
one-to-one relationship enough, or are there circumstances where you'd want
a one-to-many relationship? Other comments/suggestions? What granularity
and range should the frequency be able to represent? My draft right now
gives 100 hz steps up to about 40 GHz.
Scott / N1VG
----------------------------------------------------------------------
Subject: RE: JAVA-capable Radio
From: "Gregg G. Wonderly" <gregg@skymaster.cytetech.com>
Date: Mon, 03 Feb 2003 12:47:44 -0600
X-Message-Number: 19
>Scott:
>
>Take a look at:
>
>http://www.ibutton.com/TINI/index.html
>
>which is a embedded processor that is programmable in Java. Full tcp/ip
>stack, ethernet, 4 serial ports, one wire, digital I/O, etc.
>
>Granted, it into itself is not a complete radio/tnc combo, but it might be
>the good start for a prototype.
The Jstamp at www.jstamp.com is another example. This is the AJile aJ-80
in a DIP package with power. They have other packages with ethernet. The
prices are still high, but the quantities are still marginal.
And, as some of you may know, I've already implemented a large amount of
the Tnc infrastructure for Java based APRS/packet radio applications in my
JeAPRS project out at jeaprs.sourceforge.net.
-----
gregg@cytetech.com (Cyte Technologies Inc)
----------------------------------------------------------------------
Subject: RE: [ Miller Scott Contr 30SCS/FTI ] Re: Future spec
From: Miller Scott Contr 30SCS/FTI <Scott.Miller@vandenberg.af.mil>
Date: Mon, 3 Feb 2003 18:56:43 -0000
X-Message-Number: 20
If we had high bandwidth links everywhere, I think XML would be the way to
go. But the protocol I'm designing is supposed to be suitable for
low-bandwidth RF links as well, which pretty much rules that out. ASCII
representation of data, I'd say, is more difficult for microcontrollers
than any decent binary format. At least, it's more bloat if you're using
standard conversion functions.
Like you said, you can still fall back to byte-by-byte decoding. In fact,
I did that over the weekend, while implementing a quick-and-dirty decoder
in Visual Basic. VB is absolutely horrible for manipulating binary data,
but it can still be done. People manage to handle IP and other such binary
protocols all the time. Yeah, it's not perfect, but I for one can't think
of an easier to decode format that wouldn't be much less efficient.
Yeah, Java is great, but I happen to LIKE doing funky things with pointers.
=]
Scott / N1VG
----------------------------------------------------------------------
Subject: aprsd 2.14 no livedata at port 10151
From: Wolf Hoeller <wolf.hoeller@gmx.net>
Date: Mon, 3 Feb 2003 19:34:30 +0100
X-Message-Number: 21
Hello !
I have a 'aprsd-2.14' running with kernel-ax25 support by dl5di. When I
connect the server with XASTIR or with UI-View32 on port 10151 (live data
_with_ history) I can see the calls from the history file. But no live data
from the server will be displayed!!?? After a disconnect and a reconnect I
see the additional data - and so on..... but no live data!
Can somebody imagine, what is going wrong with my server?
--
73 de Wolf, oe7ftj
ax25-mail: oe7ftj@oe7xlr.#oe7.aut.eu
email: oe7ftj(at)gmx.at
Have a look where I am: http://map.findu.com/oe7ftj-*
----------------------------------------------------------------------
Subject: Xastir Up and Running ... Sort of
From: "Brian Riley (maillist)" <n1bq_list@wulfden.org>
Date: Mon, 03 Feb 2003 17:38:53 -0500
X-Message-Number: 22
Well I went to the website that was pointed out to me:
<http://members.chello.nl/~m.vanopzeeland/xastir.html>
This is a wonderful site. The instructions are straight forward almost key
stroke by keystroke and it all worked as it said it would.
I downloaded and installed Apple x11 and SourceForge's Fink. The used Fin to
bring down a bunch of packages (I believe M$oft's "Windows Update" team
could learn something from Fink!) The downloaded Xastir and it compiled and
setup just fine using Apples developer tools included as part of Jaguar.
Xastir comes up and runs. I configured it to my local numbers, BUT, where do
the maps go? I found a map folder in "/sw/xastir" but I cannot copy anything
to there. I have a bunch of Win/MacAPRS maps to put in there, so where do
they go/ Do I just change permissions/ Or is there something else?
I tried to configure an interface for the internet but every server I tried
seem to get an error (I had an active net connection at the time) xastir
looks pretty good but the help files are a little Spartan.
cheers ... 73 de brian, n1bq
----------------------------------------------------------------------
Subject: Re: Xastir Up and Running ... Sort of
From: "Curt Mills, WE7U" <hacker@tc.fluke.com>
Date: Mon, 3 Feb 2003 14:49:59 -0800 (PST)
X-Message-Number: 23
On Mon, 3 Feb 2003, Brian Riley (maillist) wrote:
>Xastir comes up and runs. I configured it to my local numbers, BUT, where do
>the maps go? I found a map folder in "/sw/xastir" but I cannot copy anything
>to there. I have a bunch of Win/MacAPRS maps to put in there, so where do
>they go/ Do I just change permissions/ Or is there something else?
>
> I tried to configure an interface for the internet but every server I tried
>seem to get an error (I had an active net connection at the time) xastir
>looks pretty good but the help files are a little Spartan.
On a Unix box, Xastir typically creates /usr/local/xastir and maps go into
/usr/local/xastir/maps/??
You can create any heirarchy of maps that you like below there, and even
create symlinks from there off to other areas to hold even more maps.
If the Mac OSX install puts the Xastir heirarchy at /sw/xastir, then I
would suspect your maps would go into /sw/xastir/maps/*
Please keep notes about any questions/problems/suggestions you have about
the docs (or anything else for that matter), and forward them to the xastir
mailing list. We'll make it better (or you can!).
--
Curt Mills, WE7U hacker_NO_SPAM_@tc.fluke.com
Senior Methods Engineer/SysAdmin
"Lotto: A tax on people who are bad at math!"
"Windows: Microsoft's tax on computer illiterates!" -- WE7U
"The world DOES revolve around me: I picked the coordinate system!"
----------------------------------------------------------------------
Read previous mail | Read next mail
| |