TMG-L Archives

Archiver > TMG > 2002-02 > 1012700086


From: "F. Langset" <>
Subject: Re: [TMG] Long discourse on GEDCOM and witness data
Date: Sun, 3 Feb 2002 02:35:26 +0100
References: <3C5C47B0.1830.B27A9C@localhost>
In-Reply-To: <4.2.2.20020202173803.01350e68@pop3.norton.antivirus>


-----
Ref: Lee Hoffman, 2 Feb 2002 18:00, "Re: [TMG] Long discourse on
GEDCOM ":
-----

> This discussion is open to all comers. <g>

Thank you. Actually, I hoped someone would hush me to bed. It is 2:00
on this side of the globe.

> Just for the record, TMG as currently implemented is not all that far
> removed from the GENTECH Data Model. Yes, there are differences, but
> when the GENTECH Data Model is fully accepted for use by the
> genealogical industry/community, I am sure that a new version of TMG
> will be created that fully implements the Data Model and will
> automatically convert your data to that new version.

I did not talk about the TMG data model, but the special way that I
have to arrange my data to accommodate the transfer data model
(GEDCOM today, GENTECH tom...some time, whatever in my next life).
You could say constraints to the usage of, extensions to or the data
model that I implement on top of TMG's data model (one census tag pr.
individual, no witnesses, biography tags for the witnessed mariages
etc, etc). Then, when GENTECH or whatever eventually is there, you
still have all those programs that cannot tackle such data anyway (so
I should not communicate them because I do not know how they will be
interpreted, said someone). I don't like the approach of imposing a
specific structure on my data to accommodate other peoples needs more
than my needs. So the question is: Do I have to? And is that because
of some problems of more philosophical concern or are there actually
real insoluble problems? Although I haven't read every message in
this thread down to the last punctuation, I cannot see that anybody
has shown anything that suggests real problems (terms like lost data
integrity, form of corruption, etc. etc. don't scare me unqualified
into a wheelchair ;) )

> The problem is that there is no "GEDCOM data model." GEDCOM is an
> ambiguously written "standard" based on what was understood as the
> needs of the genealogical community. That is, no data model for
> genealogy had ever been developed until the GENTECH Data Model was
> developed. However, the understanding of those needs varies from
> person to person due to the fact that there are no well-defined
> genealogical term (either in the US and (especially) internationally).
> We make use terms during conversations for which we know the meaning
> and have no problem even though the parties to the conversation have
> slightly differing meanings.

I can agree on this one, but, as I see it, there is a limited set of
possible, still usable data models. And TMG has (should have!)
selected a one of them. That is the GEDCOM data model I'm talking
about. I just don't hope that is the same one that TMG is using
internally because that one cannot be implemented in GEDCOM (at least
that is the message of this thread).

> > We must not let the holes in GEDCOM reflect into TMG.
>
> That is exactly the point that Bob is making. TMG can allow the user
> to enter data for which there is no method (within GEDCOM) to export
> the data such that the data is understood correctly by another
> program. Lacking such a method, TMG simply chooses to ignore that
> data rather than possibly (probably) export corrupt data. So TMG
> follows the GEDCOM "standard" as closely as possible -- while allowing
> the user to still enter non-exportable (via GEDCOM) data _IF_ they
> wish. Alternatively, TMG also allows the user to enter data in such a
> way that it may be exported via GEDCOM. The problem with this last is
> that some TMG capabilities are not used. The user just has to
> determine which is the most important -- using the advanced
> capabilities of TMG or not.

I do not agree that there is no such method for the types of data
that have been discussed here. I do agree that some aspects of the
data kept in TMG will be lost, but that is not the same as the data
that was exported being corrupted. I feel that one try to tell me
something like that (on a very general level) if I had a text
"1.00000000" and converted it to a real number, I have corrupted the
data because if I convert it back again I may get "0.1E-1". I do not
agree unqualified to that. Again: We must not let the holes in GEDCOM
reflect into [our data in] TMG.

IMHO, I think we can benefit from being a bit more pragmatic.

Sorry if my american/english is bad and my inhibitions are all gone,
it's hours after bedtime ยง=) g'night.
Regards,
Frode Langset
Norway
----
P.S. If this didn't make any sense: Print it out and
throw it in the waste basket, then give me a notice.

This thread: