BL>          I'm writing a PD PKT twitter for use with TinyPoint. I won't
 BL> tell you that I'm going to use Visual Basic because it will only upset
 BL> you, so we can pretend I'm using portable ISO-C compiled by Watcom.

Great to see!  Nothing wrong with using Borland though.  When you've
written in Borland, you can always get me to compile the final thing in
Watcom, because so long as you stick to 100% ISO C, you won't have any
portability problem. That's the big advantage of writing your C programs
according to the C standard.

 BL>   Just looking at the packet, it seems that every message begins with
 BL> C702 C702 0000 0000 and I can use this to identify messages. I have

02C7 in hex is 711 in decimal.  The 711 is the 711 in 3:711/934.

 BL> tried it manually, just overwriting the message with spaces, and it
 BL> seems to work okay, but I'm a worried there may be a trap for young
 BL> players (and poor old pensioners).

Yes, that's a very bad thing to do.  You are essentially screwing up the
packet by doing something like that.  And if you screw up the packet,
different programs will do different things.  Tobruk would stop processing
and tell you that your packet is screwed. Other programs just ignore that
message and try to resynch.

 BL>   I have a two questions:

 BL> 1.  Is the C702 C702 0000 0000 a reliable identifier?

No.

 BL> 2.  If not, what is?

02 00.

 BL> 3.  Will over-writing spaces between the C702... etc damage the packet?

Yes.

 BL> 4.  Where can I get the PKT specifications?

FREQ FTS-0001.* from 3:711/934.  It will answer all your questions on this.
Also FREQ "TOBRUK" and you will get the C source code to
something that reads packets.

 BL> 5.  Is true love the most important thing in the Universe? (I just threw
 BL>     that in, on the offchance you'd know).

No, it is just one of the things.  Savate is equally as important. BFN.  Paul.