Jay Sprenkle wrote:
On 7/26/06, John Stanton <[EMAIL PROTECTED]> wrote:
>
> Almost a "plug-in" serialization organization. I don't think I've ever
> seen anyone do it that way, but that might be really valuable.
> The embedded software guys would just love it.
We had an ISAM product which worked after that fashion. It replaced a
server-type implementation which in unfamiliar hands was more trouble
than a bag full of monkeys.
The isam product was barrel of monkeys fun, or the server implementation?
The only thing I can see that would be a problem would be chosing the wrong
model and ending up with corruption because you aren't serialized.
--
SqliteImporter and SqliteReplicator: Command line utilities for Sqlite
http://www.reddawn.net/~jsprenkl/Sqlite
Cthulhu Bucks!
http://www.cthulhubucks.com
The server was a constant source of problems because it had to be
installed on the host machine and integrated with the startup, not
always easy in an earlier era of departmental computers and junkyard dog
system administrators. If it stopped for some reason it generated
service calls. An ISAM manager integrated with the the applications was
much less trouble, and I can see that situation recurring with Sqlite a
fortiori.
Your point about problems due to the wrong model is very pertinent.
Perhaps there could be a check compile mode, like compiling in range
checking in a Pascal program, to intercept at run time gross
misconfigurations.
