Re: extensible DBMS with ADT of spatial type is really spatial DBMS ?

From: Paul G. Brown <paul_geoffrey_brown_at_yahoo.com>
Date: 12 Nov 2001 15:09:51 -0800
Message-ID: <57da7b56.0111121509.61dc7afb_at_posting.google.com>


waterium_at_sina.com (Neonate) wrote in message news:<67fc27d2.0110312130.6e2829e3_at_posting.google.com>...
> But i
>
> don't think this is a real spatial DBMS , because the defination of ADT is
> defined by user ,not implemented by DBMS itself, While a real spatial DBMS is
> implemented by DBMS itself. That's ,for the so called sptial DBMS based on
> extensible DBMS ,the spatial type is implemented in the level of application;
> while for the real spatial DBMS ,spatial type is implemented in the level of
> system!

  Can you come up with an application were the difference between a system supplied type and a user suplied ADT makes any difference at all?

  Being a DBMS engineer, it has always puzzled me that people seem to hold decided opinions on what "ought" to be in the DBMS and what "ought" to be somewhere else. There is no "ought". There is only what makes engineering sense. Managing spatial information, and most particularly geo-spatial and spatio-temporal information, involves using some very complex algorithms. So the DBMS engineers sensible got the specialists (ESRI, MapInfo, Geodesy) involved. Making this work, whether or not the types/behaviours are categorized as system defined or user defined, is more or less the same from the DBMS internals perspective.

  This sounds like a marketing difference. It is not an engineering difference.

  KR

          Pb Received on Tue Nov 13 2001 - 00:09:51 CET

Original text of this message