Re: Identity modelling

David Cressey wrote:
> ...The context of the thread was
> "identity modeling". As the OP of this thread
> pointed out identity modeling and relational modeling are not exactly the
> same thing. Well, actually the OP was responding
> in the dbdebunk thread, but mAsterdam turned that response into an OP when
> he changed the name to the thread. (did masterdam change the identity of the
> thread, or just its name?)

Re-modelling the NNTP distributed database, are we? I don't know much about that.
I changed the value of the subject-tag (SMTP, I suspect) but I /did/ reply to a post.
Did I start a new thread? I really don't know. I did not break the reply-chain, but I broke the subject-chain.

RFC 977 does not contain the word 'thread'.

> The word "key" has meaning in identity modeling, AFAIK. A key value
> identifies an entity. It identifies the same entity, regardless of whether
> it surfaces as foreign key or a primary key in the RM, or merely surfaces
> as a key attribute in the ER model.
>
>
> Aside from the trivial questions of official definition, which x was so kind
> as to provide us, the real question is whether
> the URL is used as if it were a pointer or as if it were a key. There's
> some schizophrenia in the IT community about this, whether the community
> wishes to recognize it as such or not.
>
> For example:
>
> If I move an html file from one directory to another on the same disk, have
> I changed its identity? Have I changed its URL? The answer to the second
> question is clearly yes. A hyperlink bound to the old URL will now return
> "Page not found".
> But the question about whether I've changed its identity by moving it is a
> more subtle one.

You say "moving /it/" (emph. mine), you moved something with an albeit unstated identity (it) from one place to another.

Why did not you copy it before deleting the original? :-) Received on Wed Aug 31 2005 - 21:04:30 CEST