Hi!
I'm posting this to the mailing-list because I get the feeling that this
thread becomes quite entertaining. If you're not, I certainly am! :-)
[Comments are welcome, to my private address, please]
On Tue, 11 Aug 1998, Thomas Michalka wrote:
> Set the following case: Somebody lives in a region where a connection is charged
> by impulses, the first one *always* with a constant duration, e.g. 10 minutes.
> The following impulses may be different depending upon what time it is and/or
> the day in the week.
You have just described my initial problem again! :-)
It is regardless which length the first (initial) impulse is of *IF* it is
different from all others.
Just a remainder: I shortended the first impulse to account for the time
during which my modem and that of my ISP's probe the analog telephone
line.
Put this another way and it reads:
"I live in a country where the first impulse is always _shortened_ by 15
(replace your number here) seconds."
So the initial impulse is *always* shorter than the rest, by a *constant*
margin. Got me?
> This construction needs a treatment by diald with an initial duration affected
> by a fuzz parameter, so both impulses, the initial *and* the second one, has to
> be affected by <fuzz>.
That is what I said in the first place. <fuzz> must be applied to all
impulses for diald to shut down the link after the first (i.e. before the
second) impulse.
> With 'impulse 600,60,20' diald has to probe the link state after 580 seconds at
> a first time and then after 640 second (from the very first beginning) a second
> time.
> So, finally I agree with you that this would be a useful behavior of diald.
Again: As useful as your example above. No difference!
> I have read the manual page once more with a maximum of attention. Again I would
> say that it tells us the fuzz parameter actually should *not* affect the initial
> duration:
>
> "In the three argument form, the <initial-duration> parameter indicates the
> minimum number of seconds diald will keep the line up once a call has been
> initiated. After this timer expires diald will proceed as for the two
> argument case using the <secondary-duration> parameter in place of the
> <duration> parameter."
>
> So, if your test turns out the opposite, this part of the manpage definitely
> *must* be wrong.
No, just incomplete. It may be important to note that the two argument
format is described right before. What is meant is probably that the three
argument form behaves like the two argument, only the first (initial)
impulse length is used *once* (...after that timer expires...).
The only question remaining is if the <fuzz> factor is applied to the
inital impulse as well. That started this thread.
**** ISN'T THERE SOMEBODY WHO KNOWS THE SOURCE ??? ****
[Sorry, but I couldn't help it...]
> > Can you explain this behaviour by reading the man-page?
> I can't too.
See, me neither. ;-)
> That's it why I didn't believe the fuzz parameter working this way (did never
> test it because the 'Deutsche Telekom' doesn't have an initial impulse duration
> different from the following impulses).
^^^^^^^^^
different... Here you say it yourself!
There isn't such in Austria too. Does anyone know a country where it
applies?
> Conclusion: this behavior could be better documented.
Yes, but I'd say that diald is already very well documented. What is
missing are more (explained) examples and a FAQ preventing people
extracting their answers from the man pages. As we discovered, these are
not always perfectly clear. Therefore, interprations may vary.
> BTW: The example for the three argument form
> impulse 600,0,0"
> is a bit unlucky, because here it is nearly not important when diald probes the
> link state, in the last 20 seconds of the first impulse or directly after this
> impulse every second.
This confirms my request for more detailled examples.
> Now I think these things are cleared, but not yet improved. This has to be done
> for the next change of version.
Does this mean you agree that the fuzz factor is applied to both?
I feel kind of sorry that this thread may has come to and end... ;-)
Many thanks, Thomas!
Regards, Walter
--
Haidinger Walter <[EMAIL PROTECTED]>
Student of Electrical Engineering, University of Technology, Vienna, Austria
Address: Brunnerstr.6, A-3108 St.P"olten, Austria. Tel.: +43-2742-257191
-
To unsubscribe from this list: send the line "unsubscribe linux-diald" in
the body of a message to [EMAIL PROTECTED]
