John,

John Clarke wrote in a message to Russell Brooks:

 RB> Here is proof that a KTX (Rockwell) internal WILL CONNECT (& Vice 
 RB> Versa) to a USR Courier at 33600 under perfect conditions.

 RB> Perfect conditions meaning a 2 metre modular cable or if both ends of
 RB> the link are sitting on their exchanges (central offices).

 JC> [snip]

 RB> So the problem lies (thats the right way to spell it in Telstra case) 
 RB> with Telstra, not the modems.

 RB> Now everyone, lets write to Austel and get rid of Telstra.
 RB> Give Optus a chance to show what they are make off.

 JC> Which will achieve exactly nothing.  Optus will have similar
 JC> guidelines regarding local loop lengths (i.e. loss) - in fact while
 JC> they are setting up they will likely be worse than Telstra.  One
 JC> exchange serving all of Melbourne, as was reported here in the last
 JC> week or so, doesn't set the conditions for short local loops.

this is waht all the optus chorus don't understand .. on the other hand i
will be one of the few, very few that will be most heartily amused when
optus ``turn's'' on thier loyal most voacal supporters in the bbs
community. its like my grandmother said all those long years ago, to
paraphrase, you get what you pay for.

 JC> The moral of your story is get a house as close as possible to the
 JC> exchange (Telstra or Optus, it makes no difference) and no further
 JC> than 1Km as the crow flies.  Alternatively, move to a small (under
 JC> 1Km radius) country town with a digital exchange.  Either way you
 JC> reduce the local loop loss as much as possible which means that
 JC> whoever you connect to will probably get a 33.6 connection. 

i live in a sydney seaside suburb, within spiting distance of the open
sewer called clovelly beach or was that coogee .. any way one of them. i
have been having a interesting time getting my microcom deskporte fast+ to
work over the copper we have in the ground. i'm about 1 km (as the cable
layers measure these things) from the main injection point at coogee
exchange. my connections with other deskportes are not too bad, regularly
at 28k8 but sometimes a split 28k8 with 26k4 incoming ... on the bad days
its 24k incoming.

if you add a maestro to the other end, if it connects, usually the rate
starts at 26k4 and retrain till the data gets sent or untill the microcom
gives it upo a a bad joke. this is particularly a problem with the older,
super executive 14k4, dataplex style metal cases .. the newer ones in the
plastic stand up on one edge 14k4's while better in that they hang on
longer suffer from being an order of magnitude fussier in making the
original connection.

these are the observations i've made over the 4 years i've lived in
randwick and surrounds area. also i can reproduce all of these wih the pair
of maestro hat i used  run before i  he microcom.

 JC> Of course, you will also have to convince your ISP/BBSs that you
 JC> call to do likewise to reduce THEIR local loop loss if you want
 JC> 33.6Kb/s in YOUR direction. 

one isp, used to run netcomm exclusively in thier early years and the
nightmare saga .. i heard all the excuses under the sun ...

my comms card wasn't up to scratch, they supplied one of thier super duper
several hundred dollar ns16550afn driven multiple interrup thingies, no
improvement.

my modem wasnt up to scratch .. they supplied a range of netcomm product
that had progrssively worse connectiviity untill i get feedup with the
whole exercise .. it got to the point where the m11f couldeven negotiate a
14k4 connection, if it did manage to connect it would last 7 to 15 minutes
and give up. at thet point i'd replace the it with my trusty maestro nd it
would soldier on as the advert says.

then came the desperate gropings of tech support that didn't have a clue
about what it is doing .. they latched on to the fact that i was running a
dos based uucp package, while talking to thier unix (sco and later linux, i
believe) and thierfore legitimate uucp .. funnily enought those cowboys now
headup the sydney linux user group .. i hope they display more curtocy and
professional development at thier hobby than they did while on thier day
job. 

anyway, at the end of the day i changed isp, and got far better connects,
when i do have problems, well my current providor at least knows his unix
and we can quickly get on to the carrier for further processing, of the
line faults.

i suppose what i've just outlined is not so important when it comes to
specifically modem related faults .. but, the acknowlegegement that the
modem is but one link in the whole trin of events that constitute a
successfull connection and subsequent data retrieval.

it is important to know and understand the chain, as well as having a good
working relationship with your services providor .. be they the common
carrier, the tech support at the ip injection site or you local kiddie
running its bbs for fun and or profit.

untill that state of relative nirvana is reached modems no matter how good
they get will be just another source of frustration.

 JC> If either modem doesn't support split baud rates then BOTH would
 JC> have to apply to get 33.6 connections.

just one more source of connection failure, one that is not very well
understood buy a significant number of the ever increasing modem using
public.

regards ... jonathan

EMail: jlm{at}caamora.com.au

---