Home -> Usenet -> comp.databases.theory -> Re: domains aren't subtypes, right?

Re: domains aren't subtypes, right?

From: Jonathan Leffler <jleffler_at_earthlink.net>
Date: Thu, 05 Jun 2003 04:29:13 GMT
Message-ID: <3EDEC78B.7070806_at_earthlink.net> 






Mikito Harakiri wrote:


> "Paul G. Brown" <paul_geoffrey_brown_at_yahoo.com> wrote in message


> news:57da7b56.0306031727.209ae842_at_posting.google.com...

> 

>>   Well, now we're into a philosphical discussion entirely off-topic to
>>  c.d.t. (But I share your sense of taste in this instance . . . .)




> 

> 

> OK, closer to the topic.


> 

> Proposition. The only date type allowed stored in the RDBMS should be

> INSTANCE/TIMESTAMP


> 

> Other datatypes are functions:

> 

> DATE = FLOOR(TIMESTAMP / Day)

> TIME = TIMESTAMP (mod) Day

> 

> Storing timestamp in inferior units is possible, but just a handicap to the

> system evolution and progress. Besides, what is the TIME: instant, interval,

> or just a number in a fancy base system?





I disagree.  You need to be able to store intervals (or durations), 
and those are distinctly different from points on the timeline (which 
is what a timestamp represents).  Now, if you said the only type to be 
used was an interval, with timestamps simply being an interval from a 
fixed point on the timeline, then you might be in with a chance - but 
not a very good one - of justifying your proposition.


-- 
Jonathan Leffler                   #include <disclaimer.h>
Email: jleffler_at_earthlink.net, jleffler_at_us.ibm.com
Guardian of DBD::Informix v2003.04 -- http://dbi.perl.org/


Received on Thu Jun 05 2003 - 06:29:13 CEST












Original text of this message