TMG-L Archives

Archiver > TMG > 2003-12 > 1070337548


From: Dennis Lee Bieber <>
Subject: Re: [TMG] Wishlist for TMG5
Date: Mon, 01 Dec 2003 19:59:08 -0800
References: <000201c3b7aa$8c15ad20$d8feadcb@main><000201c3b7aa$8c15ad20$d8feadcb@main>
In-Reply-To: <3FCAA0C5.9040507@mindspring.com>


On or about 18:00 11/30/03 a carrier pigeon from David Madeo delivered:

> 2) GPS: Allow different formats for data entry. Convert if you
> must. Many of us have GPS's which use WGS84.

ALL GPS receivers use WGS84 -- that is the native gravitational
model and reference for the GPS constellation. It, however, has very little
to do with the coordinates you enter. For coordinates the normal choices
are Lat/Long (usually in variants of DDD.dddddd, DDD MM.mmm, DDD MM
SS.s), Universal Transverse Mercator (or Military Grid Reference System).
UTM/MGRS uses slices of the globe, and measurements in meters east from the
slice reference, and northward.

On that basis I would like to see TMG accept UTM/MGRS coordinates,
because they allow for easy "mapping" of related points (point Y is n
meters east of point X and m meters north, assuming both are in the same
"slice").

On top of this, and maybe of more import, is that most maps are
based upon some reference datum; any given coordinate is a different
location if you change the reference datum. Most GPS units default to the
NAD-72 (I believe), but most USGS maps are still based on NAD-34 (or some
similar year). Some newer maps do have markings for the delta between the two.

And if one were to be recording coordinates for the UK, one should
probably use both OSGB coordinates (if the GPS unit supports them) AND the
OSGB datum.

Essentially, the coordinates are telling you /how/ to measure
positions from some reference, and the datum is telling you where the
reference is.



> 6) XML output. I haven't looked into whether there are any good
> schemas, but I use XML for EVERYTYHING at work. Genealogical data
> would make sense as well.

I believe the LDS has been working on this for GEDCOM 6.x --
though the last time I looked at it, the spec seemed backwards to me for
parsing purposes... It defined "family" first, with indexes to the
individuals, and then had the individual data listed... Any easy conversion
of this to a database that enforces referential integrity would require one
to first parse the individuals so that the family foreign keys (to
individuals) are valid.

--
> ========================================= <
> | Dennis Lee Bieber <
> PGP keys available | RSA keys preferred <
> ========================================= <
> Home Page: http://www.dm.net/~wulfraed/ <


This thread: