IS> Looking at it from the point of view of Laplink-like function, you're
 IS> right, but Telix and other comms programs work just fine via a
 IS> properly setup null-modem cable, to another instance of themselves or
 IS> to BBS programs.  Using Zmodem (say) it's fairly trivial to transfer

 FF> I've never had much luck using Telix through a null modem cable.  Tried
 FF> 'blind dialling', locking comm ports speed, etc.  Kept falling over after
 FF> a few seconds so I assumed Telix just wants to use a real modem.

 FF> Could you post the relevant Telix settings in this echo please?  This
 FF> technique will be of widespread interest I think.

Long gone - but nothing special, as I recall.  The only real trick is
getting your RTS/CTS handshaking working properly via the null-modem cable,
and having it enabled in the comms program too, of course.  RTS 
CTS (4  5, DB-25) should do the trick.  I always get good results
connecting DTR to the other end's DCD, too, so Alt-H 'drops carrier' the
other end (DTR hangup enabled).

 FF> I should have been more specific :)
 FF> 
 FF> The 4.77 MHz IBM-XT with its standard 8250 UART serial port
 FF> won't reliably
 FF> go faster than 9600 baud under INTERLNK/FILELINK.  These
 FF> programs apparently
 FF> do not perform any error-checking and rely on the hardware
 FF> working perfectly.
 FF> 

Fairy nuff.

 IS> In 1990 I transferred my entire system (all 20Mb :) from a 10MHz XT
 IS> clone to my then new 386sx16 - before the 386 had its 16550AF UART -
 IS> at 57600bps using a shareware 2-node network called WeakLink, with
 IS> almost no slowdowns due to CRC retries.  In fact it was only just too
 IS> flaky at 115200bps.

 FF> I've used FILELINK at 19200bps to swap stuff between two 10MHz XT clones,
 FF> both fitted with 8250s, but 38400 transfers kept falling over after
 FF> 20-30 seconds.  All it takes is *one* error.....

Seems to say more about FILELINK than the UARTs :)

 IS> The XT did have an 8250A (eqivalent in to the 16450 on an AT), but
 IS> 38400bps on 8250s is easy enough going - the only limitation to
 IS> 19200bps was the standard INT14h DOS functions, which no smart comms
 IS> program went anywhere near ..

 FF> But can INTERLNK/FILELINK be considered 'smart'?  They
 FF> don't do error-checking
 FF> for one thing.  I suspect they both use INT14h and (egad!) X-modem packet
 FF> transmission.

Doubleplus unsmart .. there are a zillion shareware gadgets out there that
do better than that.  One out of many I tested that went ok was ZIP, but
the aforementioned WeakLink (by the author of the $25 Network) was really
solid at speed, with reliable error-correction (using RTS/CTS h/shaking),
and was just installed as more or less a master/slave network at boot,
adding remote drives so you could use your favourite utilities to move
files - I was just using XTREE at the time.  Even worked ok in Desqview,
rather to my amazement ..

 IS> ATs and slow 386s (single tasked) haven't the slightest problem at
 IS> 115200bps on null modem.  A 486 won't give you any advantage for such
 IS> =slow= transfers as 115200bps over, say, a 12MHz AT.  Even

 FF> As I said before,
 FF> 
 FF> are you ready to reveal This Knowledge about Fast Telix Null Modeming to
 FF> the teeming masses of Ignorant Mankind?
 FF>        :)

I'm trying to remember, it's been quite a while - I use Binkleyterm for
just about everything these days - but I really don't think I did anything
special at all; just connected the null-modem as if it were a real one :) 
I think Telix might need DSR to be true on startup, but don't quote me on
that .. I even remember using some existing dial directory entry, seeing
the modem init strings pop out the other end.  You may need CTS blocked
during disk writes.

 FF> And thanks for the info; I stand corrected.

Well the XT only had a 1.2Mb floppy, and I was too lazy to copy it all over
that way .. so I had to fiddle till it worked - I was surprised, too!

Cheers, Ian

--- MaltEd 1.0.b5