Hi Luca!

> ...
> a)
> every insert operation should use the id generator method
> that uses the table above ;^)
> If you add new records into a table "by hand", you could increment the max id
> value for all the tables without update the "high" field value of the
> "app_identifier" table.
> So there is the possibility to create an insert statement with a duplicate key 8^(
> 
Yes, that's true, but at least there is a possibility to coordinate
several independent applications which all respect such a mechanism. 

Maybe the best would be to be able to switch from one algorithm/method
to another by changing a simple attribute within config file. One of these
methods could be "databaseMaintainedGenerated". We could
also consider to define an interface that can be used to plug in own
implementations. As an example, I am using a dbms that uses a similar
approach like your one with the separate table within applications.
To be able to run these old applications together with the new dbforms
applications, I _must_ use the current tables, so the possiblity 
to plug in an own implementation would be great. Until now I have to 
use interceptors. 

Another problem might be that if you use a database table, you might
face the problem of coordinating concurrent accesses. For one dbms you 
might use transactions, for another you might have to do a explicit
'lock table' command. What does your implementation do here?

Regards

Dirk


-------------------------------------------------------
This SF.net email is sponsored by: ApacheCon, November 18-21 in
Las Vegas (supported by COMDEX), the only Apache event to be
fully supported by the ASF. http://www.apachecon.com
_______________________________________________
DbForms Mailing List

http://www.wap-force.net/dbforms

Reply via email to