Subject | Re: [IB-Architect] RE: [Firebird-devel] New system functions/objects was: planning t he use of blr codes |
---|---|
Author | Jim Starkey |
Post date | 2001-04-24T16:50:05Z |
At 12:12 PM 4/19/01 -0400, Leyne, Sean wrote:
about Interbase/Firebird was implemented in violation of the SQL
standard. This includes dates, blobs, multiple transactions,
triggers, two phase commit, etc.
Sean, if you let other people do your thinking you are doomed
to a permanent position on the trailing edge of technology.
The original Interbase philosophy was "SQL -- don't fix it,
don't extend it, just implemented it." SQL as prevailed
because they wanted something called a standard more than
they wanted an actual standard. So now the mantra should
be "SQL -- no arbitrary differences." If desired functionality
can be incorporated within the bounds of the published standard,
it should be. If there is good reason to diverge, do it
carefully.
A security model that recognizes that a server is an agent
of a client is required for progress. I don't expect the
SQL committee to recognize this for another five years. I
don't propose we wait until then. Do you?
Jim Starkey
>I would like to point out that absolutely everything that people like
>As Diane will probably point out, SQL does not allow for multiple user
>roles.
>
about Interbase/Firebird was implemented in violation of the SQL
standard. This includes dates, blobs, multiple transactions,
triggers, two phase commit, etc.
Sean, if you let other people do your thinking you are doomed
to a permanent position on the trailing edge of technology.
The original Interbase philosophy was "SQL -- don't fix it,
don't extend it, just implemented it." SQL as prevailed
because they wanted something called a standard more than
they wanted an actual standard. So now the mantra should
be "SQL -- no arbitrary differences." If desired functionality
can be incorporated within the bounds of the published standard,
it should be. If there is good reason to diverge, do it
carefully.
A security model that recognizes that a server is an agent
of a client is required for progress. I don't expect the
SQL committee to recognize this for another five years. I
don't propose we wait until then. Do you?
Jim Starkey