Re: Oracle's client internals

From: Michael Winters <>
Date: 13 Dec 1993 12:04:28 -0500
Message-ID: <2ei7as$>

In article <>, Erwin Dondorp <> wrote:
>I have no intention of rewriting any Oracle client tool completelly,
>but I was wondering about the following question:
>- Can every single oracle client product potentially be rewritten
> from scratch by a skilled person with Pro*C ?
>I find at least the following problems:
>- How does a client know in the case of a syntax error at what
> position this error is? (does it do the parsing also?)
>- How does a client know the version string that it prints
> after you succesfully logged in?
> In mean the text:
> Connected to: ORACLE7 Server Release - Production
> PL/SQL Release - Production
> This is information that the client gets from the server!,
> how can user written clients get these strings?
>- How does a client decide whether it is sending a SELECT statement
> or an other statement? (does it look at the text? other?)
>Erwin Dondorp Xirion bv, Software and Consultancy
> Burg. Verderlaan 15X
>Currently at work at: 3454 PE De Meern
> PTT-T-NWB-NWO-PCS-SDC the Netherlands
> 070-3433613 Phone: +31 (0)3406-61990 Fax: 61981
>Een heer van stand is geen nummer! (OBB)

I think that someone skilled in Pro*C could make more efficient software, but it would obviously, take a lot of effort, in some cases. I have a related question to this - If someone is looking to maximize program efficiency, shouldn't they use OCI calls, where you are accessing the Oracle Kernel directly, without Pro*C having to precompile the calls into the individual Oracle database functions? I haven't ever worked at a lower level than Pro*COBOL or Pro*C, in an Oracle environment, but had a Systems Programmer ask me about "accessing Oracle via database kernel calls", so I am curious about this subject, too.

  • Mike Winters U.S. Peace Corps
Received on Mon Dec 13 1993 - 18:04:28 CET

Original text of this message