Re: Commitment Control (was: Modernizing applications (was: Explaining single level store tonon ipeople))

From: Charles Wilt
What happens if the program is ended via user action, system
crash, power failure, ect... between the write to one table and
the update of the other?

The time between a write to one table, and an update of another? We're talking about millionths of a second. Ended via user action? How many users act with that degree of precision? Power failure? Use battery backup. Lightning strikes the server? Well, then your problems are bigger than a missing update to an inventory record.

You've got two choices
1) Use commitment control
2) Add lots of extra code into your application for recovery.

I have no idea what you mean by #2.

In this day and age, why would you write develop and application
without commitment control?

We can talk about that after somebody explains why people develop and deploy Adobe Flex or Microsoft Silverlight applications to workstations that accumulated transactions on PCs, then posts them as batches to .Net or PHP application servers, which interface with MS SQL Server databases on a 3rd Machines, then take issue with what might fail in the millionths of a second that it takes for an RPG program using RLA to lock, write, then update. ;-)

-Nathan.