TMG-L Archives

Archiver > TMG > 2002-02 > 1012765874


From: "Darrell A. Martin" <>
Subject: Re: [TMG] Having it both ways (was: Long discourse...)
Date: Sun, 03 Feb 2002 13:51:14 -0600
References: <3C5D63F1.20828.1165350@localhost><4.2.2.20020202224222.01313d60@pop3.norton.antivirus><3C5CA1EE.1391.F22174@localhost>
In-Reply-To: <4.2.2.20020203114439.013f8f00@pop3.norton.antivirus>


At 12:01 PM 2/3/02 -0500, Lee Hoffman wrote:
[snip]
>However, you can have it both ways. This requires redundant entry of
>data. You would enter Witnesses in the normal TMG Witnesses section of a
>tag. Then you would enter the data in a different tag (could be a Custom
>tag) and mark the Tag Sentence of that tag such that it will never print
>and/or de-select that tag in the Report Definition. Then when you
>generated a GEDCOM, you would specifically select that tag for export.
>
>
>----------
>Lee Hoffman/KY

Hi, Lee:

I tested this method. For the results, see my separate message. I hope
someone finds my research useful. However, I believe the method is fatally
flawed.

If every time I enter a census tag witness who is not a principal I
remember (a non-fatal flaw <grin>) to also add my custom GEDexp tag for
GEDCOM export, I have to decide what text to put in GEDexp's memo field. My
choice is to use the output of the census tag witness sentence. That, by
the definition implicit in the construction of TMG, is at least one
acceptable representation for output of the data I have entered. The census
tag witness sentence is part of the context in which I entered the data;
the TMG program is the environment that gives my data entry meaning.
(Reductio ad absurdam on the "just the way you entered it" front: where in
a current TMG-generated GEDCOM do you see, "first I clicked the big + sign
on the square yellow button"?)

Anyway, say six months from now I figure out that the modification of the
default tag witness sentence that I had put in place before starting all
these new shenanigans with GEDCOM export, stinks. So I go back to the
default. (The other way around has the same problem.) For all my printed
reports from that point on, the change affects all my census tag witnesses.
But for my GEDexp tags? I have to find each and every one, separately, and
manually modify it. Unless I have created a separate GEDexp equivalent for
every tag type, I will have to look at every witness tag in the entire data
set. And typos *will* occur, roughly 1 every 500 keystrokes. I will catch
about two thirds of those.

Storing redundant data is not corrupting the data. I have written many
programs that do that. But if no automated mechanism exists to audit the
correspondence of the duplicate copies, corruption of the data is inevitable.

Nope. Back to ripping out the non-principal witness tags. Which I'd be half
done with, by now, if you hadn't side-tracked me with such a promising
suggestion. Shame on you <big grin>.

Darrell


Darrell A. Martin
a native Vermonter currently in exile in Addison, Illinois



This thread: