> A better client/server approach is to populate
> the grids with a single or small dataset as a
> "selector" and, in the same transaction, to
> perform another query to fetch back the single
> row for editing...this model is commonly termed
> "drill-down" and IBO is highly oriented to it.

Yes. Nonetheless, on smaller networks such optimizations
are often not necessary so that many fields can be
queried for many rows at once without anyone complaining
about insufficient speeds.

At least nowadays, this is often possible!

Cheers,
Tobias