Jonathan Leffler wrote:
> paul c wrote:
....
>> I don't remember this aspect of that model ever being discussed here
>> but I contend it is just as reasonable as Codd's original intent,
>> which was in part to make access to data storage more intuitive.
>
> I wonder if part of the issue is the name? As I understand it, the TRM
> is a storage level device, operating as a suitable substrata for a fully
> relational DBMS to use - but not itself constrained by all the rules of
> the RM. For example, it unabashedly has columns in a defined order, not
> in an undefined order.
> ...
I think you are right about that, maybe my long-windedness made it look
otherwise. As I read the descriptions of trm and parts of the patents
it looked to me that sometimes it doesn't define the order of columns
and sometimes it does. As a user that would annoy me. I would want the
order of result columns to be consistent.
> So, it is a possible physical level representation of data that the RM
> handles at the logical level with the rules of the RM running the show
> at the logical level.
>
> When were the patents on the Trans-Relational Model issued, does anyone
> remember? I wonder how much longer before they expire - I expect it is
> a decade away before those in the USA can use them unfettered, but it
> might be a bit less.
>
> I'm also not clear whether Codd ever really discussed or got involved
> with it. C J Date included an appendix in his 8th Edn, and apparently
> wrote a book on it but that got stranded when the company ran out of
> money and the VCs and founders no longer saw eye-to-eye, or thereabouts.
>
> OK - maybe my new subject for the thread is as misleading as the old. It
> is still "Trans-Relational Model Moribundus", but the topic of
> discussing it has been revivified.
>
|
