Geoff,

> > Yes, perhaps even at design time! We could make it have
> > all the pros of persistent fields without the quirks
> > and shortcomings. Setting an event property at design
> > time would merely have the function of reminding the
> > component to do an assignment at run time, after prepare.
> > A much less invasive thing than persistent fields.
>
> This line of thought went through my mind - and I think that is what
> scared me off. There are lots of issues and things could get quite
> messy (integration with existing Fields* properties, the SchemaCache
> and more).

I see; I was not suggesting to move the settings from the dataset to the
field objects (after all the work Jason has done to go the exactly
opposite route), but rather make available at design time what currently
isn't, perhaps a couple of events *only*.

> I suspect that **if** it does happen will remain a runtime only apsect
> at least for IBO4 (anyway thats my opinion). Probably such a change
> would remain largely invisible to the normal user and be setup
> primarily to make some aspects (such as assignment of event handlers)
> more intuitive.

That would be a benefit, at least for the question that originated this
thread. I believe, however, that design time support (again, limited to
event handlers, which should do no mess) would be a big plus,
particularly for many newcomers to native IBO that don't initially get
to grips with the absence of persistent field objects to link event
handlers to (oncalculate and onvalidate being the most commonly used).

I have appreciated that it's one of the most difficult things to explain
when I had to present IBO in a Delphi/IB class.

Let's wait for Jason to read and ponder on this...
Ciao
--
____
_/\/ando