Columbia Kermit/kermit.zip/newsgroups/misc.19950528-19950726/[e-mail]_Sat Jun 17 18:41:46 1995.msg

Internet Message Format | 2020-01-01 | 6KB

Received: from apakabar.cc.columbia.edu by watsun.cc.columbia.edu with SMTP id AA08189
  (5.65c+CU/IDA-1.4.4/HLK for <[e-mail]>); Sat, 17 Jun 1995 22:14:13 -0400
Received: by apakabar.cc.columbia.edu id AA13993
  (5.65c+CU/IDA-1.4.4/HLK for kermit.misc@watsun); Sat, 17 Jun 1995 22:14:12 -0400
Path: news.columbia.edu!sol.ctr.columbia.edu!howland.reston.ans.net!news.sprintlink.net!demon!not-for-mail
From: Edward Doolittle <[e-mail]>
Newsgroups: comp.protocols.kermit.misc
Subject: Robust kermit dialing scripts (was Re: kermit 3.13 question)
Date: 17 Jun 1995 19:41:46 +0100
Organization: Demon Internet News Service
Lines: 94
Sender: [e-mail]
Message-Id: <[e-mail]>
Nntp-Posting-Host: dispatch.demon.co.uk
X-Posting-Host: brauer.math.toronto.edu
Apparently-To: [e-mail]

Newsgroups: comp.protocols.kermit.misc
Expires:
References: <3qne5b$[e-mail]> <3qq847$[e-mail]>
Sender:
Followup-To:
Distribution:
Organization: Department of Mathematics, University of Toronto
Subject: Re: kermit 3.13 question
Summary:
Keywords:

[Apologies for this late response... news posting has been out at
our site for several weeks, and it took me some time to find an
alternate site from which to post.]

> In article <3qne5b$[e-mail]>,
> Jeff Schneider  <[e-mail]> wrote:

>>    1. how can I invoke kemrit with a batch file so that it will either:
>>        a. connect directly trough a com port
>>        b. dial the modem - area code optional,
>>                    prefix optional
>>        c. connect directly through a network card.

In article <3qq847$[e-mail]>,
Frank da Cruz <[e-mail]> wrote:

> Still, your script program can't be a mind reader.  It doesn't know if the
> laptop it is running on is in the owner's house, in a hotel, on an airplane,
> or plugged in to a network (or if it is, whether the network should be used
> instead of the serial port or internal modem).  So there will still need
> to be some hints from the user.  That is, the user will still have to know
> something.  Like whether s/he is in a hotel room or at home.  And what
> the dialing prefix is, and whether an area code is needed, etc.

The script could try to connect over the network. If failure, it could
try the serial port, sending CR or LF or BREAK and waiting for the
ogin: sequence. If that doesn't occur, then the script could try the
modem.

Probably the most reliable way to dial the number is to have the user
dial the number on an attached telephone and have the local modem take
over once the remote modem answers. That would probably require a modem
with both line and phone jacks, but it is a robust method for making the
call, permitting the user to deal with any human telephone operators
encountered, and with any complicated local instructions for making a
call (e.g., "dial 9 then the long distance carrier identifier,", etc.).

That the user has to know something beyond the number to which they want
to connect is a bug, not in kermit but in the phone systems that we
have constructed.  If I am on the Interenet and want to connect to a
remote site, all I need to know is the IP address of the site. Why
should the phone system be any different?

For example, if I want to dial a local number, I can't use the 1- long
distance prefix.  What kind of design decision is that? That means that
I must know my current area code as well as the destination area code
(and in some cases, even the distance from source to destination!) in
order to call anyone. Or I have to use trial and error. That's easy
enough for most humans most of the time, but what about machines?

A similar example appeared in the Risks Digest (comp.risks) recently.
An empty pop machine kept calling an office desk phone number. It turned
out that the pop machine had been moved from one area code to another,
so that the number that originally belonged to the supplier's computer
became the number of a human being in an office. If the area code could
have been included in the phone number, that would not have occurred.

Problems like these could be solved if we pressured the phone companies
enough to do something about them. In a time of increased competition
from the Internet and cable companies, I'm sure they'd be delighted with
suggestions for improving the utility of their systems.

>                                                             Yes, the don't
> have to be helpless in the face of complicated technology.  What is the 90's
> word for that...  empowerment?

Doing it once is empowerment. Doing it 100 times is just slavery once
again.  Yes, I could talk directly to my modem or figure out the
required dialing prefix by consulting all of the provided reference
tomes, but once I know what to say to it I want to give the job to a
script. Real empowerment is creating a machine to do the job for you.

The real problem here is that we assume too much of users, not too
little.  The hugeness of our assumptions only becomes apparent when we
try to replace some of the functions performed by users with functions
performed by machines.

Complexity in itself is not a virtue. While successfully handling
complexity may make us feel good, unsuccessfully handling it, or
handling it over and over again, may cause quite the opposite effect,
and may in fact lead to unanticipated dangers.

Ed