Re: [BUGS] BUG #14842: TIMESTAMP WITH TIME ZONE and TIMESTAMP WITHOUTTIMEZONE should not be comparable

On Tue, Oct 3, 2017 at 2:19 PM, Eduardo Pérez Ureta <edpeur@gmail.com> wrote:

I did not know that the SQL standard was so broken. Maybe it is time to consider only following the SQL standard in its non-broken parts, like other SQL databases do.

​It has its quirks but the two you have fixated on seem easy enough to avoid.

At this point we've already decided that being conforming is generally a virtue; especially since I'd assume that at least some "other SQL databases" also conform to these same rules.  Deciding to become more strict (as opposed to starting out that way) requires significant benefit.

IMHO, at *most* postgres could issue warnings on usage of sketchy constructs.     I think emphasizing the warning in the documentation is sufficient, however.

-- 
john r pierce, recycling bits in santa cruz