WriteLog
[Top] [All Lists]

[WriteLog] "double numbers" and out-of-sequence problems

To: <writelog@contesting.com>
Subject: [WriteLog] "double numbers" and out-of-sequence problems
From: Marty Tippin" <martyt@pobox.com (Marty Tippin)
Date: Thu, 25 Jan 2001 21:40:11 -0600
As I have mentioned to Wayne privately before, I don't think the issue of
gaps in serial numbers, or out-of-order QSOs in the log is nearly as much of
a problem as DUPLICATE SERIAL NUMBERS.

I'll venture to guess that the log checkers (either human or computer) don't
really care if every number in the sequence is represented, or if they're
represented strictly sequentially in your log file.

However, when QSOs in two other logs claim you gave them the SAME serial
number (which the log checkers will find), somebody is going get PENALIZED
(either the person who sent the duplicate number or one of the two who got
it) -- and it's all due to an ERROR in the software they're using. That's
absolutely unacceptable, especially when there's a way to prevent it!

A relatively simple solution to the duplicate serial number problem is to
have a chunk of code that does nothing but hand out serial numbers, and
which ensures that the same number is never handed out twice (i.e. a
"critical section" for the multi-threaded programmers among us.)

I'm a professional Windows/C++/COM/Java/Database programmer and I deal with
this stuff all the time, and I can assure everyone that the solution is not
that difficult and that it will work, with guaranteed results.

This solution will allow gaps in serial number sequences (for those times
when the serial number is discarded because the QSO isn't logged) and will
certainly allow QSOs to be logged with serial numbers out of sequence (which
doesn't matter in the least), but will NOT (EVER!) GIVE OUT THE SAME SERIAL
NUMBER TWICE.

Wayne - can you please tell me why this kind of solution can't be
implemented in WriteLog? I'll be happy to lend whatever assistance you'd
need in getting it done - I'll crank out a thread-safe
serial-number-generating COM object if you need me to, and you can have the
code to do what you want with.

Looking forward to a reply,

-Marty NW0L
 martyt@pobox.com

----- Original Message -----
From: "W. Wright, W5XD" <w5xd@writelog.com>
To: <k9ig@contesting.com>; <Writelog@contesting.com>
Sent: Thursday, January 25, 2001 9:01 PM
Subject: [WriteLog] "double numbers" and out-of-sequence problems


>
> The short answer is: no, it has not been corrected, and in addition,
> I am not working on anything to change WL's current behavior.
>
> I am sure that's not a very satisfying answer, so here is more
> explanation.
>
> The bottom line is I cannot fix what you are calling a problem,
> because it is impossible to run two radios in a serial number
> contest without having "flaws" in the serial numbers.
>
> The easiest scenario that proves this is to consider running S&P
> on both radios and you start a QSO on radio A by dropping in
> your call. He sends his exchange to you, you respond by sending
> yours to him, and, at the instant your transmission on A ends,
> you drop your call in on radio B. If the guy on A asks for a
> fill (or never QSL's at all), and the QSO on radio B finishes,
> (with a little juggling by you because you can only transmit
> on one radio at a time). Then you will either get an out-of-order
> (A eventually finishes) or a gap (A never finishes).
>
> I admit though, that WL does allow more gaps and out-of-order
> that it "needs to". It could reduce these by assigning the
> serial numbers later in the QSO--when they are sent. But
> WL doesn't really know when you send the serial number,
> espcially for SSB. WL also allows duplicates, which I could
> make never happen (until you add more PC's on a network--then
> I can't stop duplicates either).
>
> The reason I have no plans to do any more work on this is that,
> because of the above scenario, EVERY single serial number
> contest weekend is going to be followed by postings on this
> reflector saying the same thing the attached posting says:
> "It's still broken". Yes, sir, it is.
>
> So why bother? I have other things to work on that I can
> actually succeed at fixing. And who has had QSO's disqualfied
> because of out-of-order, duplicates or gaps? (Folks HAVE
> lost QSOs because of mis-logged numbers, but WL does
> guarantee it logs the number as-sent).
>
> (My shields are now up. Fire away)
>
> Wayne, W5XD
>
>
> -----Original Message-----
> From: owner-writelog@contesting.com
> [mailto:owner-writelog@contesting.com]On Behalf Of Greg Clark
> Sent: Thursday, January 25, 2001 12:43
> To: Writelog@contesting.com
> Subject: Re: [WriteLog] confusion about upgrades
>
>
>
> Has the problem with QSO numbers (double numbers, out of sequence) been
> corrected
> in this new release? It's not stated in the release notes and I've
received
> no direct replies from Ron.
>
> Is it still broken?
>
> Greg
> K9IG
>
>
>
>
> --
> WWW:                      http://www.writelog.com/
> Submissions:              writelog@contesting.com
> Administrative requests:  writelog-REQUEST@contesting.com
> Problems:                 owner-writelog@contesting.com
>
>
>


--
WWW:                      http://www.writelog.com/
Submissions:              writelog@contesting.com
Administrative requests:  writelog-REQUEST@contesting.com
Problems:                 owner-writelog@contesting.com


<Prev in Thread] Current Thread [Next in Thread>