Re: Why use OCI when there's Pro*C ?
Date: Wed, 3 Mar 1993 20:59:33 GMT
Message-ID: <1993Mar3.205933.7014_at_csis.dit.csiro.au>
In article <C30G2J.Avv_at_jbecpor.demon.co.uk>, smj_at_jbecpor.demon.co.uk (Martin Jarvis) writes:
|> Why do people use C and OCI when they could use Pro*C ?
|> What can you do with OCI that you cant do in Pro*C ?
|>
|> These are questions I have asked many times but can find no answer.
I find that Pro*C does not map nicely onto the OO paradigm. For example cursors are declared statically in Pro*C. It does not seem possible to write Pro*C code which creates a cursor and returns a pointer to it or another which destroys a cursor given a pointer to it. This is possible in OCI.
We are writing a data server which currently uses oracle at the back-end. Because a data server can simultaneously have many clients, using Pro*C would enforce a fixed number of possible cursors and thus a fixed possible number of possible server clients.
This is not a limitation that we wish to impose.
I hear that Oracle V7 is multi-threaded. I don't know if/when Oracle will support multi-threaded clients but when it does - I'm sure that static cursors won't be the go.
I don't know what is going to happen with Ingres as it doesn't even have an OCI interface. How many other RDBMS vendors do?
The Pro*C interface leaves me wondering "Why on earth did they do it like that?".
-- +-----------------------------------------------+--------------------------+ | Dave Campbell (dave_at_csis.dit.csiro.au) | |\ ____|\ | | Phone: + 61 6 275 0944 Fax: + 61 6 257 1052 | |\___ /\ \ / / \ | | CSIRO Division of Information Technology | | _ \ / \ \_/ / ____> | | Centre for Spatial Information Systems | | |> > __ \ / \ \___ | | GPO BOX 664 _--_|\ | |__ /_/ \ \_/ \___/ | | CANBERRA ACT 2601 / \ | \/ \/ | | AUSTRALIA \_.--._/ <- Canberra | Get right, or get left! | | v | Heb 2:3 | +-----------------------------------------------+--------------------------+Received on Wed Mar 03 1993 - 21:59:33 CET