On Wed, Apr 23, 2008 at 6:38 PM, Martin Gainty <mgainty@xxxxxxxxxxx> wrote: > > > Emiliano and Mike > > The real challenge is trying to determine what a datatype is in cobol..for > that matter what is stack variable or heap in Cobol? > In the end you're better off <re>writing this mess (preferably in Java).. > unless of course you need the billable hours for > the first rewrite to C > then later rewrite to Java (have no idea how this relates to the OP's original question). If you are trying to port a cobol app to postgres, your best bet is to go through the client interface, libpq. If you had to do it on the server side, I would stick to cobol environments that are C ABI compatible. Writing general purpose data procedures in C is just not a very good idea most of the time...it's difficult and dangerous...C SPI has great uses, it's just not for everything. I personally think cobol is better suited for data processing type problems than java. Mapping cobol data types to SQL is not terribly difficult. cobol is notoriously difficult to port to another langauges...probably cheaper to connect it to the database via ISAM wrapper if the app is over a certain size. Many modern cobol environments support external data sources through various techniques...extfh for example. AcuCobol (crypticly) allows linking a ISAM emulation layer directly to the cobol runtime, one approach I've used in the past. merlin
