TMG-L Archives

Archiver > TMG > 2003-01 > 1041729079

From: "Darrell A. Martin" <>
Subject: [TMG] TMG sort dates in general
Date: Sat, 04 Jan 2003 19:11:19 -0600
References: <><>
In-Reply-To: <>

At 10:57 AM 1/4/03 -0500, Lee Hoffman wrote:
>>>What I mean is-what would be the difference between "After 28 Feb 1959" and
>>>"On or after 01 Mar 1959"?
>>>Or are they the same?
>>In my opinion, they are or ought to be *exactly* the same in meaning, as
>>long as you are talking about a structured date.
>I would disagree. There is almost 24 hours difference. What time of day
>did the event occur? If it occurred (or might have occurred) at at some
>time during 28 Feb 1959, then that day is included in the "After". Now if
>you use the same specific date for both statements (After 28 Feb 1959 and
>On or after 28 Feb 1959), then I would agree that there is no difference.
>Lee Hoffman/KY

Hi, Lee:

The way you have described it is the way that TMG currently works, and I
understand that quite well. It is, in my studied opinion, simply not the
way the language works or that almost any non-TMG user would understand
things. Outside the TMG community, "28 Feb 1959 at noon" is NOT a possible
answer to the question, "name a date and time that is after 28 Feb 1959",
and it still would not be a possible answer if the time were 11:59 p.m. But
TMG thinks it is; Linda's suggestion tries to deal with this perceived
problem, another problem being how TMG sorts comparative dates of differing

I think TMG's fundamental error is in not providing for a difference in
meaning between "after" and "on or after", which it then compounds by
defining "after" as if it *meant* "on or after". It does not mean the same
thing, any more than "greater than" means the same thing as "equal to or
greater than".

It is true that if an event occurs "on" a specific date, and another event
is known to have occurred "after" the first, it may still have occurred on
the same date. An example would be if the only source document about a
woman's death said, "John Smith was born on the last day of February, 1959.
His mother died a short time later, while holding him for the first time."
Perhaps she died before midnight, perhaps not; all we can say for certain
is that she died "after the birth event". However, this cannot be used as a
defense of TMG's approach. Although John Smith's birth, to the greatest
precision provided in a TMG structured date, is "28 Feb 1959", it does not
follow that Mrs. Smith died "after 28 Feb 1959". Rather, she died "on or
after 28 Feb 1959". Accuracy will require further explanation, but this is
a separate issue from the structured date.

At least, it will be a separate issue unless TMG implements sorting tags
directly as an option. TMG does have two methods of ordering the births of
siblings: ordering by the sort date of the primary tag in the Birth Group
(I think); or, by entering the birth order directly. This is the closest
that TMG comes currently to providing for the relative sorting of events.

If any of this is to be changed, I would guess it will be by the addition
of one or more options such as the one Linda has suggested. The current way
TMG works does not appear to Bob V. to be a problem, it seems; he has often
defended it. I might also add that if an option *is* to be added, adding it
should not IMHO be a current priority. All of us who find the current
sorting to be counterproductive have either found workarounds or have
learned to grin and bear it. Well, at least bear it ... [grin].


Darrell Allen MARTIN
a native Vermonter currently in exile in Addison, Illinois

This thread: