Robert Haas <robertmhaas@gmail.com> writes:
> FWIW, I don't really buy this argument. I can't see that it's all
> that implausible to think that the user might be able to prognosticate
> a reasonable value for a future TOAST table.
Well, it still seems to me that such a user is really more interested in
a way to set the default toast fillfactor (or whatever option is under
discussion), ie what he really knows is a reasonable value for *all*
future TOAST tables in his installation.
Otherwise you're arguing that he knows exactly what the fillfactor
should be for a specific toast table and not any other one ... except
he doesn't know when that toast table is going to be created, which
calls into question the quality of his judgment about its specific
behavior otherwise.
regards, tom lane
