Re: [Gramps-users] Re: [Gramps-devel] Upcoming enhancements

[Howard H. <hh...@ho...>]

Interesting, I'm quoted below, but nothing is quoted :-)

In any case, if I understand the request, I think both The Master
Genealogist (TMG) and _the_most_recent_versions_ of Family Tree Maker
(FTM) both will adequately handle the simple case of conflicting data
you give.  A genealogy program must allow conflicting data to be of any
use.  Not just for multiple users, but if a birth certificate has one
date and a family bible a different date, those two dates both need to
be recorded along with the source information so that it's possible to
track down errors over time and to draw the best conclusions based on
ALL the data available to the researcher.

To better motivate the shared data problem in my earlier message let me
espouse a more complex example.  (And I appologize for the rather hard
to follow message last night.  I wanted to get the ideas out sooner
rather than later, but it was late and I wasn't at my most lucid.)

Let's say I'm working on my family history and have a file of direct
ancestors of myself and my wife.  A cousin is also working on family
history and has a file of direct ancestors of himself and his wife.
Obviously, there is some overlap, we both have ancestors of (let's say)
my maternal grandfather.  Now, he gives me his file, I notice that he
has a given name and birth date purportedly from my grandfather's birth
certificate.  I have a copy of the birth certificate and see that
clearly he has two typos in his file.  I update both the name and the
birth date as recorded on the birth certificate.

Exercise for the reader.  What do I give back to him, and how does he
know what to fix?  Keep in mind, in the real world there are probably
several fixes including maybe a branch of the ancestors he knew about
that I didn't and one that I knew about that he didn't.  If I give him
"my file," the next time we do an update, he already has my wife's
family tree and all the updates I've made may conflict.  Note that he
made no changes to that subtree and doesn't really have any interest in
it.  If I give him back his file corrected, he may not find and/or make
all the corrections that I think are appropriate.  Next time he sends me
the file, his "updates" may wipe out corrections I've already made even
in data I am actively tracking and taking responsibility for (my
grandfather's birth date, remember?)  Various combinations of merged
subtrees have similar problems.  The problems expand as the pool of data
sharing users expands (e.g. my cousin, my mother, my sisters, and a few
long lost relatives who have "complete" genealogies of limited
reliability.) That's the problem "multiple database views" and "uniquely
identified records" were designed to solve. Whether my solution
completely solves it is up for debate, of course. It's a very hard
problem, and it's hard to recognize all the complex situations and make
sure that "something reasonable" happens.  The grandfather example here
is particularly nasty in that a name and birth date change will often
convince genealogy programs "merging" data to completely duplicate whole
sections of the family tree because the two records are clearly
"different people" to the program.

On 02 May 2001 08:55:08 +0200, Bo Rosen wrote:
> Den 02 May 2001 01:13:43 -0400 skrev Howard Holm:
> 
> Let me first say that I am not a programmer, so please take any comments 
> with that in mind. I don't know what is possible or practical.
> 
> My top most priority is proper internationalization, this includes character 
> sets, translations, data entry formats (dates for example), paper sizes.
> 
> The ability for different people to work independently of each other and add 
> data to a common data base is something that would be very useful, on
> different computers or centrally. I'm sure anyone who does genealogical
> research together with someone else and on more that one computer would
> dearly love this. There is the problem of "syncing" the data, what if I
> and my mother both enter a date for a person, but have different dates
> for the same event? Best solution would be to add both dates, but set
> some type of flag, a warning that the record contains duplicate dates
> for an event and needs to be checked. Is this doable?
> 
> I understand that gramps.data uses xml, does this mean that in theory
> you could do all sorts of advanced searches and analyses, via plugins,
> sripting or some other way?  Say I want to find all women born between
> 1758 and 1878 named Maria (including variants, Mari, Marie, etc) who
> were married more than once and didn't die giving birth.
> 
> Cheers,
>     Bo
> 
> 
> 
> _______________________________________________
> Gramps-users mailing list
> Gra...@li...
> http://lists.sourceforge.net/lists/listinfo/gramps-users
> 

-- 
Howard Holm  (hh...@ho...)
(GPG Public Key etc): <http://www.holmgrown.com/people/hholm/>