TMG-L Archives

Archiver > TMG > 2002-02 > 1012596641


From: Bob Velke <>
Subject: Re: [TMG] Witness and Roles and V5.0
Date: Fri, 01 Feb 2002 15:50:41 -0500
In-Reply-To: <Springmail.105.1012591652.0.90606200@www.springmail.com>


Darrel said:

>However, I believe there is an assumption in your
>response, above, that is false. That is, that if
>the TMG event to which one is a witness is called
>a "baptism", then the export function in TMG must
>call any and all data generated in GEDCOM format
>from that TMG event, a "baptism". TMG can, and
>does with ease, distinguish sentence texts
>between principals and witnesses. I do not see
>why there should be any problem with an equivalent
>kind of distinction during GEDCOM generation.

Well, then, you haven't read the GEDCOM specs lately <g>. GEDCOM is not a
blank slate on which I can do anything that I want. (Some developers treat
it that way but that doesn't make it legitimate GEDCOM). There are
distinctions that I can make in a GEDCOM file and others that I cannot.

>So the question is not, for me, *whether*
>valuable information in TMG should be "permitted"
>transfer into a GEDCOM file, but rather *how* to
>best do so with the least misrepresentation of
>the data in the process. Doing so does not require
>maintaining the integrity of the data right down
>to the format it is stored and handled in, or the
>structure of its linkages. It is much like
>language translation; one strives for dynamic
>equivalence, not for a rigid one-to-one
>correspondence.

The difference between one-to-one correspondence and a translation is that
the latter requires the communicator to _interpret_ the data and that is
exactly what GEDCOM transfers are intended to avoid. And it is not
intended to communicate a _report_ or reformulation of the data by a
programmer. The effort is to faithfully preserve and communicate raw data
exactly as it was entered by the researcher and that requires minimizing
the number of times that the data is converted. That precludes
constructing narratives and using other formulas to convert the data from
one format to another before transferring it.

If you don't care if witnesses, etc., are exported in a _structured_ way
(that is, a narrative will suffice) then you can do that already. You can
generate a narrative report with TMG and then cut and paste it in its
entirety into the memo field of a NARRATIVE tag or something of the
kind. By doing so, you accept responsibility for it as your own data
(instead of a programmer's construction) and then you can export it to
anything that can read a GEDCOM file.

-Bob

This thread: