> Date: Tue, 27 Apr 1999 09:20:36 -0400
> From: Patrick Ouellette <[EMAIL PROTECTED]>
> Subject: Compression in packet radio
>
> I've watched the discussion about the need for compression to be added
> to the various services run over packet radio. Why is everyone so
> intent on modifying *every* service individually? What we need is to
> have the ax.25 protocol layer negotiate compression on connect. Then
> individual users and sysops can decide if they want to allow compression
> (or require it) on their systems. If the transport protocol is doing the
> compression, the individual services don't have to be modified - it would be
> transparent to them. Let's put the effort in creating a solid compression
> strategy where it belongs, and every program can benefit from it.
Based on a casual inspection of the APRS protocols, I think that:
o We need to do a better job of applications-level protocol
design. Personally, I think a series of binary-encoded
type/length/value (TLV) tuples is a much better encoding
than ASCII text strings. (It seems like Part 97 permits
binary protocols in the U.S., but I don't know the regulatory
implications for the rest of the world.)
(TLVs also provide a nice mechanism for extending the protocol,
because receivers can skip TLVs they don't implement.)
o An AX.25-level compression protocol may be useful, but probably
can't make up for poorly applications-level encoding.
For example, a lower-level compression protocol can't determine
what bits are simply irrelevant and should be tossed, (e.g.,
fixed-location decimal points, longer-than-necessary keywords,
etc.).
-tjs
