Re: Overloading words: How can we get Oracle to stop it?
- From: "Ghassan Salem" <salem.ghassan@xxxxxxxxx>
- To: dbvision@xxxxxxxxxxxx
- Date: Sat, 10 Jun 2006 17:38:13 +0200
Guys, when I see this, I remember a quarrel about the recylebin, when it
appeared. Should the dropped tables appear in user_tables or not? how the
table in recylebin should be named, .... IMHO, it all comes down to backward
compatibility, and how to (or not to) preserve it. In the case of
dba_tab_privs, most probably (I cannot check), when it appeared (may be in
7.0), it had only the tables in it. When other objects came along, Oracle
had to either add a new table, or use this one. If you use the same, can you
change the column names? it will raise an avalanche of requests to revert
back as it may break lots of software, if you add a new table, the same
software will have to be modified to use this new one, ....
The poor guys (probably they're not poor) at Oracle Labs will have a hard
time on such things.
As for the initial email about the 'partition' word. Well, it's meaning is
adequate in both cases (i.e. table/index partition, and in analytical
functions), and I've seen lots of Oracle Clients use both without ever
seeing any problem with this 'reuse'. Also, the analytical functions spec
were written with IBM's cooperation, so it must have not been easy to choose
the words for the different clauses.
just my 2cents
On 6/10/06, Nuno Souto <dbvision@xxxxxxxxxxxx> wrote:
Mark W. Farnham wrote,on my timestamp of 10/06/2006 4:02 AM:
>
> And yes, I'm tossing in a huge presumption here that anyone else
> actually cares.
>
Some of us do. My pet peeve is dba_tab_privs and its columns
"TABLE_NAME" for a table and a column respectively that store
*object* names! Ie, suddenly package, type, sequence and
view names are *table names*????
--
Cheers
Nuno Souto
in sunny Sydney, Australia
dbvision@xxxxxxxxxxxx
--
//www.freelists.org/webpage/oracle-l
Other related posts:
