On 2011 Apr 28, at 5:52, Corn Walker wrote:

> On Apr 27, 2011, at 11:36 PM, Richard S. Russell wrote:
> 
>> I was musing in the car on the drive home tonight that one approach to databases is to have the data stored in one file and the user interface in another. This got me to speculating on the prospects of having a multi-file system where all of the SCRIPTS are in a file of their own. Has anybody ever tried this? How did it work out?
> 
> It doesn't work out very well. Scripts execute based on the context of the file they are in, so unless they're all in your UI file, having them elsewhere is a non-starter.
> 
> Separating the data and the interface into separate files does work out, and it's something people do all the time. Placing all of the scripts into your UI file will work with some caveats, the primary one being that security is defined at the file where the data is defined. So if you want to prevent a user from deleting records except with your script, and you use the security restrictions to do this, you'll need a script in the data file to override those security restrictions.


Thanks, Corn. This saves me a bit of what would have proved to be fruitless experimentation.

<rant>Which, of course, reminds me of one of my pet peeves: File-Drawer Syndrome. This is where some researcher advances a hypothesis and experiments to test it, only to find that it's disconfirmed. Rather than publish that finding, tho, the researcher drops the study into a file drawer somewhere, never to see the light of day. Meanwhile, a dozen other researchers have the same bright idea and set about testing it independently. Of course, they'll end up with the same negative results and will probably not publish those, either. What a waste of time and brainpower! Negative results are results, too! We need to pay more attention to them. </rant>