Re: Genealogical evidence and data model

Dennis Lee Bieber <wlfraed@[EMAIL PROTECTED]
> writes:

> > of change in state. The "event" in TMG apparently remains tied to
static
> > data. Is any kind of event allowed besides the obvious ones of birth,
> > marriage and death?
> >
>	What would you like?

Don't pay too attention to my remark, which percipitated from my primary
concern for historiography and the "event" in that context. In that
context, it is in principle possible to escape a static view, but I'm
not sure about genealogy, which to some extent is primarily concerned
with data.

>> No, I wasn't thinking of a heap. A heap is an array or tree structure
>> that orders people according to lineage. It seems very old fa****oned
>
> 	I'd call that a hierarchical file structure... or simply a tree. At
> least when I took comp.sci., a heap file was just that -- an unsorted
> stack of dis-similar records; one had to read the record in order to
> even figure out what the data represented.

Interesting. We are appparently speaking of the same thing. When I took
my Data Structure course, we encounted the "heap" as (quote from my
textbook).  

  A heap is a binary tree with these characteristics:
  
    * It's complete. This means it is completely filled in...
    * It's (usually) implemented as an array. ...
    * Each node in a heap satisfies a heap condition, which states that
      the node's key is larger than (or equal to) the keys of its
      children.  
  
In my notes on databases, I have:

  A heap is a specialized tree-based data structure that satisfies the
  heap property: if B is a child node of A, then key(A) ≥ key(B).

As you point out, there is no suggestion that record B has any
similarity to record A here. The only condition is that the key value
for record A must be at least as great as the value of the key of record
B.

This struck me as somewhat descriptive of a family hierarchy in that
there are generations, especially today when any biological relation of
families or generations is unclear, and even surnames can be only
loosely related to biological descent. For example, only one of my 28
grandchildren and 5 great-grandchildren are actually my biological
descendents because I live in a culture in which "family" has little to
do with biology and more to do with a political decision (much like
"tribe" used to be). And yet there seem to be clear generational
demarcations (as long as we can forget the etymological meaning of
"generation").

So I have to take back my comment about heap tree being suited to
lineage. It may better suit social relations today that are not so
lineage-based.

>> Incidentally, what kind of data structure is used in TMG? Do all
>> genealogical applications use the same data structure?
>>
> 	TMG is built upon Visual FoxPro tables. Very many of them --
> approximately 29 separate tables.

Interesting; I remember FoxPro from back in the 90s. Is Microsoft
dropping sup****t for it because VFP runs under wine? How will that
affect TMG? It sounds like it is a relational database that has a table
interface. Is that so?  I assume it is more than spreadsheets.

> 	None of the genealogy programs I'm familiar with can "modify"
> data /in/ a raw GEDCOM. They all perform batch loads into a database
> schema -- where data is not a sequential text format -- manipulate the
> data in that format, store the data in that database, etc.

OK, as I suspected. LifeLines does this as well, although it looks as
though one were modifying raw GEDCOM.

-- 
 
       Haines Brown, KB1GRM