Re: Extending my question. Was: The relational model and relational
Date: Sun, 23 Feb 2003 22:45:25 +0000
In message <%EX5a.232$hl5.34162743_at_mantis.golden.net>, Bob Badour
>"Bernard Peek" <bap_at_shrdlu.com> wrote in message
>> In message <7Ik5a.164$8e.17186852_at_mantis.golden.net>, Bob Badour
>> <bbadour_at_golden.net> writes
>> >It's interesting how you are trying to mischaracterize Date as using
>> >primitive technology. Multisets are demonstrably more primitive than
>> >relations--simply from the fact that they rely on physical location. You
>> >the one promoting the use of "Roman numerals and shells".
>> Multisets may rely on physical location, although that's a special case.
>That's a very remarkable statement. By what other means may users of a dbms
>using a multiset logical data model make use of duplicates?
Whatever works. What they need is a way to select one of the duplicates. What that is will depend on whatever other attributes the duplicates have. There is no inherent requirement for that to be physical location.
>> The general case is that they rely on information outside the database
>> to distinguish different items.
>Really? That's quite extraordinary. How does the dbms use this external
>information in its logical data model?
>> If that happens to be physical location
>> then you can add the current latitude, longitude and altitude as a key.
>Are you being facetious?
>> There are some practical problems but in theory it's possible. In
>> practise if you need to distinguish two cans of tuna you create an
>> identity field, call it "Can Number" and print labels to stick on each
>I've identified plenty of cans on shelves, but I've never seen a "Can
>Number". Where might I find one of those?
-- Bernard Peek bap_at_shrdlu.com www.diversebooks.com: SF & Computing book reviews and more..... In search of cognoscentiReceived on Sun Feb 23 2003 - 23:45:25 CET