🚀 go-pugleaf

We have used C++ Builder 3 to natively talk to a Progress database running on a
Unix platform.  We were using Paradox and Paradox's locking symantics on a
Novell server but ran into record contention issues we could not solve.  We were
constantly deleting .lck files.  We then took a look at Interbase on Linux but
gave up due to lack of documentation -- maybe it has gotten better but seemed to
be a strong solution.

Finally, we wound up creating our own db API and use it to talk to a Progress
broker on Unix through use of Progress software's client networking for Win32
PCs.

Another very good tool is CodeBase.  Codebase now has some decent gui tools we
may look at getting back into.  Originally our team left Codebase due to lack of
gui support early on.  Some of our DOS applications used a Codebase (x-base)
library to connect 300 some people to an order entry system for a telemarketing
company and performance was very good for a $600 database.  CodeBase will work
with dbIV III, FoxPro, client server on Unix.  We were using it on a Novell
server with record locking trained by the app.

To answer your question, find native means to talk to your databases first.
ODBC is very network chatty... We see noticable benefits between ODBC and native
drivers time and time again.


Jerwynn Lee wrote:

> Hi,
>
> We need to build a middleware system that will process huge amounts of data
> coming from
> Oracle and Sybase databases.  The program will transform these data by
> performing
> conversions and computations, and will dump the output to a new database.
>
> Is BC suitable for this kind of a program?  How would you rate / compare the
> power of
> BC when it comes to database functionality? How would you compare it to
> Visual C++, and Delphi?
>
> What kind of database access does BC support (e.g. ODBC, embedded SQL)?
> Do you still have to deal w/ the IDE layer? What API's are available for
> Sybase and Oracle?
> How is the speed?
>
> Thanks.
>
> Jerwynn Lee