Re: Creating dds source from SQL Tables

Rob,
 
To answer your question:
 
"I did not say PRIMARILY.  I simply said it was an extension."   
 
I understood you to infer that having record level access as well as
access via SQL was something that made the DB2/400 a better or at least
as good an RDBMS as the other industry standard RDBMSs, like the real
DB2s and ORACLE.
 
"And I don't know why having that extension honks you off.
 
Are no other databases an "industry standard RDBMS" if they have any
extensions? "
 
Well, first, "extensions" are usually NOT to the access architecture of
the RDBMS, only to the SQL language.  If a vendor has a different RDBMS
architecture, they don't call it an "extension", they call it what it
is... an RDBMS architecture/function/feature.   But, what "honks" me
off, is the fact that record level access, is verboten in any well
recognized, industry standard RDBMS; it's a non-standard BACK DOOR to
them.   Further, the i5 community continues to show it is not living in
reality if it thinks it can keep these legacy "let's try to please
everyone and be hugely backward compatible" architectures and, at the
same time, be seen by the rest of the RDBMS industry standard community
(the real DB2s and ORACLE) as an RDBMS to be taken seriously, in the
same league, or a competitor, or a viable purchase consideration.  It is
an accepted STANDARD that all serious RDBMSs can only have their data
accessed via the SQL language.  
 
And this is only one example where DB2/400 differs so much from what is
considered an industry standard RDBMS. Other examples of these are: the
name DB2/400 and not just DB2 indicates it is not of the same ilk as the
real DB2s,  DB2 IDUG doesn't seem to recognize DB2/400 as a real DB2 as
it has never had much to offer at its conventions about DB2/400 and how
it is truly one of the DB2 family,  RPG as an access language, talk of
Physical Files and Logical Files instead of only the architecture and
terms of TABLE and VIEW, using some other methodology and architecture
of garnering info on tables and columns other than in the RDBMS  "system
tables" such as the real DB2's SYSIBM.SYSTABLES, or ORACLES equivalent;
just to name a few.  The i5 community can't seem to understand that its
platform is considered non-standard, weird, legacy, and a host of other
names that don't bode well for its long term survival and thereby
overshadowing its potential and wonderful capabilities.   And by touting
 "weird" traits, like record level access and not exclusive use of SQL,
you only make the platform's demise much more surer and quicker.
 
Sincerely,
 
David Odom