hi Ann,

> It's hard to come up with really good rules about indexes. The normal
> example of a field that should not be indexed is sex - where there are
> two values and about equal distribution. On the other hand, if you
> had a table of inhabitants of a nunnery, an index on sex would be good
> if you often wanted to find the males.

OK. I undertand it. thanks Kjell too. So an Index can be a bad index
now, but in long term it can be a good index after thousands thousand
insertions (for example, a date index in a sales table).

Today (and yesterday) I´ve downloaded some soft for index analysis. It
generally take this as bad indexes:
nodes/totalDup >= 0.90
nodes/maxdup >= 0.30

Well I think this will cause problems in old versions but not in FB
2.x (at least for GC).

>(What's the difference between theory and practice? In theory,
> there is no difference.)

I and all our staff LOL reading your quote. It is true.

Thank´s again for share your knowledge.

Elton