Csaba Nagy wrote:
2d) Hints will damage the ongoing development of the optimizer by
reducing or eliminating test cases for its improvement.
You have no evidence for this.
My evidence (which I think I've mentioned in a couple of previous
postings), is the experience with the optimizer of that... err.. other
database that has hints, plus the experience of that (different) other
database that does not allow them :-) Several others have posted similar
comments.
2f) Hints may demoralize the developer community - many of whom will
have been attracted to Postgres precisely because this was a realm where
crude solutions were discouraged.
I still don't get it why are you so against hints. Hints are a crude
solution only if you design them to be like that... otherwise they are
just yet another tool to get the work done, preferably now.
I fail to see why would be a "hinted" postgres an inferior product...
A rushed. and crude implementation will make it an inferior product -
now not every hint advocate is demanding them to be like that, but the
tone of many of the messages is "I need hints because they can help me
*now*, whereas optimizer improvements will take too long...". That
sounds to me like a quick fix. I think if we provide hint-like
functionality it must be *part of* our optimizer improvement plan, not
instead of it!
Now I may have come on a bit strong about this - and apologies if that's
the case, but one of the things that attracted me to Postgres originally
was the community attitude of "doing things properly or sensibly", I
think it would be a great loss - for the product, not just for me - if
that changes to something more like "doing things quickly".
best wishes
Mark
