With IBO I do this way:
a) I discover a bug or an area that can be improved
b) I try to discover, in the newsgroup, if it's really a bug or I'm
missing something (of course, some time the bug is evident). In the
case of improvements, I ask in the newsgroup if my idea are right and
if there are further idea that would be good to implement (often only
Jason answers me :()
c) I patch the code for my own testing and use, trying to find if
there are collateral effects.
d) I send the source modified to Jason, that puts it "on the queue"
and tests my patch himself if theoretical appropriate.
e) I find the patch included on the next one or two sub release.
As I mentioned in a previous post, I'm very very satisfied the way
Jason incorporates suggestions/patches (not only by me, of course!),
especially if compared with Borland attitude.
At first I thought that Jason had to "work for please my requests" and
so I was a little disappointed from the fact that he was not
immediately doing what I reported/suggested ;) (sorry, I'm not
English, so I don't know if I've been able do explain the idea).
Then I discovered that he has to concentrate on "the real things", I
mean low level problem and complex bugs that only his experience and
knowledge can solve, but he really appreciate if you submit him
already made patch and improvements (he included 3 of mine in the last
10 days!).
You can't ask him "add this 'x' at the end of line 100", he has no
time to do so, just send to him the full source with the patch and his
"source confrontation utility" will notify/include the changes for
him.
I also discovered that including patches is not so easy. When I fixed
a problem on the lookup and sent it to Jason, 3 day after I discovered
that, in another situation, my fix caused other side effects, worse
than the bug I fixed, so we/he must be very carefully!
So patch the source, include it (the full source) in a mail explaining
him the problem and the fix, send the mail and wait with faith :)
Regards
Marco Menardi
Btw, I use native IB_xxx components, so I have no idea if your
suggestion is good or not, sorry.

--- In IBObjects@y..., "IBM T" <vatt@i...> wrote:
> Hi,
>
> The TIBODataset throws an exception when we have a call of the
InitFields
> with design-time fields which aren't in the table (ie.
client-calculated
> fields, lookup fields). The code:
>
>
> procedure TIBODataset.InitFields;
> ...
[... snip ]]]
>
> HTH