[Zope-dev] case insensitive sorts

[Casey Duncan]

Andy Dawkins wrote:
> 
> > Dieter Maurer wrote:
> > >
> > > Chris Withers writes:
> > >  > Dieter Maurer wrote:
> > >  > > Andy McKay writes:
> > >  > >  > .... what does anyone else think....
> > >  > >
> > >  > > I would not like it.
> > >  >
> > >  > Why not? ;-)
> > > I would not like to see this sort order in the management
> > > screens, because I use capitalization to ensure that
> > > essential objects are at the top of the object list.
> >
> > Hmmm... that's like saying you'd rather not have a memory leak fixed
> > because it gives you an excuse to buy more RAM ;-)
[snip]
> 
> I would have said its like saying your not going to fix the hole in your
> water pipe because you use it to fill up your kettle without getting out of
> bed, and if you fixed it then you would have to walk to the sink.
[snip]

I have to agree with Dieter here that a useful side affect is lost if
the sort was case insensitive. I myself found it a bit strange at first,
but now I am using different naming rules for different objects so they
mingle together. (i.e. Interface level objects get caps, logic level
objects don't).

Your analogies imply that this behavior is a bug or an unintended flaw
in the design. I would argue that it is intentional. Unix file systems
work the same way. Try doing an "ls" with mixed case files and you'll
see what I mean.

Anyhow if this goes away, I think the management interface would need to
have ability to sort by something other than id to compensate. To a
certain degree a sort by meta type would be similar for me, but
certainly less flexible, and wouldn't help compensate for the scaling
problems of the ZMI.

It just goes to show how even obscure and subtle features like this can
be important (to some at least). At least now you have the flexibility
to not use this "feature" by using the same naming conventions for
everything.
-- 
| Casey Duncan
| Kaivo, Inc.
| cduncan@kaivo.com
`------------------>