Re: Software implementing Gentech Genealogical Data Model

Bob Velke wrote:
Ian said:

How do you represent this? There is no repository. Even if one were to designate the cemetery as the repository it no longer contains the evidence.

That is irrelevant. The objective is to document where you got the evidence, not the present location of that evidence.

You could consider each and every copy of the publication as the top level of the hierarchy.

Unless you consulted each and every copy of the publication, they are not your source.

I'd suggest that one objective is to tell anyone else where they can see the evidence for themselves in which case in which case any copy will suffice. The term "each and every" was considered. *Each* copy is a source in its own right but one would give a reference to the edition as a whole, which covers *every* copy.

In any event your objections do not deal with my point: there are sources which don't fit nicely into the concept of a repository at the top of the hierarchy. And the model doesn't even need it. The rest of the hierarchy is the source entity/class which has a pointer to a higher level source - just the way to implement a hierarchy whilst remaining open-ended as to the number of levels. All you need is to accept that the pointer to the higher source can be null where it's the top level and you don't need a separate entity/class to cap the hierarchy.

In other respects it's not bold enough. For instance the UML labels some objects with a UID (unique identifier) to replace the primary database key of the original model: "A UID is only required to be unique amongst the instances in its class". A moment's thought shows that this isn't likely to be true.

You are quoting from Stanley Mitchell's reference model in UML, not the GDM itself. Keys are used in the GDM in order to construct a logical model. As an abstraction, the GDM deliberately avoids describing a physical model or implementation, including the system or authority for unique identifiers that a developer might use.

The GDM has a discussion about what makes a good key. What I suggest is that there would have been considerable benefit in a bit of lateral thinking here. I suspect that the GDM has been influenced, consciously or otherwise, by RDBMS approach and in particular data types such as SERIAL in Informix or Identity in Sequel Server. I've used this as an approach to glue systems together in the past. However in this case I'd suggest that there's merit in thinking a little more about the meaning of "unique". Surely one of the benefits of using a standard model is that data from the model can be shared. And this means that there should be some means of ensuring that "unique" means "unique" across all databases.

--
Ian

Hotmail is for spammers. Real mail address is igoddard
at nildram co uk
.

Relevant Pages

  • Re: Salvation
    ... >newsgroup is to point out objections to the beliefs held by ... >My reason, whether given by Nature or by God, seems to work ... whenever you could have been inclined either to see evidence ... prejudice. ...
    (soc.religion.christian)
  • Re: Can an individual seek a declaratory judgement of non-infringement?
    ... There is no rule requiring an attorney to be involved. ... prove that all documents submitted as evidence are authentic, ... respond to authentication objections in a manner that gets the objection ... cross examine a witness to prove he is a liar and doesn't know much, ...
    (misc.legal)
  • Re: Proposal - a new strategy to counter anti-evolutionists
    ... >>> I feel you've missed the entire point of my objections. ... the radiocarbon dating results on the Shroud ... >> paranormal proponent to provide evidence (and otherwise to disdain the ... > In other words - show me a vampire then we'll consider the question again. ...
    (talk.origins)
  • Re: What is a "nutritional expert," exactly?
    ... detail on the studies we posted) you should not post in a *.sci newsgroup. ... than trying to understand what the evidence as a whole is suggesting, ... A distinctive fatty acid profile in circulating lipids of Korean ... GDM lowers DHEA and Arachidonate. ...
    (sci.med.nutrition)
  • Re: The ToE is not dying, it was never really "alive."
    ... posted by Bob Casanova: ... Fossil evidence corresponding to plate tectonics, magnetic striping, ... He didn't mention DNA or the genetic hierarchy ...
    (talk.origins)
Loading