Re: model inherited object
From: Erwin <e.smout_at_myonline.be>
Date: 23 Jun 2006 02:37:03 -0700
Message-ID: <1151055423.880583.321480_at_m73g2000cwd.googlegroups.com>
Date: 23 Jun 2006 02:37:03 -0700
Message-ID: <1151055423.880583.321480_at_m73g2000cwd.googlegroups.com>
> Aaahh = but how do you record the "unchange"
How to record it is not the problem. How to determine that some change is actually some "un-change", on the other hand, is. I'd say, accept the idea that a database is an inventory of our beliefs about the world. Possibly about the history of the world. If those beliefs change, the database (content) must change. If those beliefs have changed and then revert to the beliefs as they were prior to the change, then that simply is a new change of beliefs and thus simply a new change of database content. Forget about "un-changing", "annulations", "un-doing" or however you want to name it. It's a mess even without the temporal dimension. Received on Fri Jun 23 2006 - 11:37:03 CEST