Re: [Gramps-users] Fw: Re: REFN

[Benny M. <ben...@gm...>]

2014-02-20 0:49 GMT+01:00 Enno Borgsteede <enn...@gm...>:

>  Benny,
>
>
>   REFN was designed to be preserved, meaning that no program will change
>> it by itself.
>>
>
>  Yes, but different people can just give the same number, and if you
> merge, your efforts of using REFN are wasted anyway. So it's a bit of a
> false safety. The gramps id is exported as the gedcom handle, so it is
> present in the gedcom. Then it depends on the other programs how they work
> with it.
>
> Well, I can tell you about those. All program that I know on Windows use
> integers, which means that where possible they strip the I from Gramps
> persons IDs, and put it back on export, removing leading zeroes on the way.
> IDs that can't be read like that are re-assigned, which means that
> customizations are guaranteed to be lost. I can see this in RootsMagic,
> where I can even inspect the person table, because it's sqlite.
>
>   It would be easy to code that on export the gramps ID needs to be
> stored as REFN. An xml parser could do that too: add a REFN attribute in
> the xml identical to the gramps id. You could do that before exporting to
> gedcom, and would have the benefits of seeing the grampsid (your REFN)
> everywhere.
>
> H'm, what about an existing REFN attribute then?
>

Gramps allows duplicate attributes, so for Gramps it would mean a person
which has more than one REFN assigned.

Benny

>
> regards,
>
> Enno
>
>